id stringlengths 36 36 | url stringclasses 504
values | title stringclasses 475
values | section stringlengths 0 148 | content_type stringclasses 3
values | has_code bool 2
classes | text stringlengths 147 22.5k | vector listlengths 384 384 |
|---|---|---|---|---|---|---|---|
4ac4f698-0dc2-5e82-a4b8-a2197fe3c529 | /documentation/tutorials-build-essentials/video-anomaly-edge-part-1/ | Video Anomaly Detection Part 1: Architecture, Twelve Labs, and NVIDIA VSS | Connecting with NVIDIA VSS > Video Chunking for VSS | documentation | true | Video Anomaly Detection Part 1: Architecture, Twelve Labs, and NVIDIA VSS > Connecting with NVIDIA VSS > Video Chunking for VSS
The first step in the VSS pipeline is chunking. Following the pattern from the [Twelve Labs x NVIDIA VSS manufacturing sample](https://github.com/qdrant/twelvelabs-nvidia-vss), we split video... | [
0.028827093541622162,
-0.05313289165496826,
-0.07151081413030624,
-0.03289419785141945,
0.054421842098236084,
-0.008902485482394695,
0.04026144742965698,
0.0203647930175066,
0.08420353382825851,
-0.03472660109400749,
0.017440982162952423,
-0.011889738962054253,
0.08104009181261063,
-0.0389... |
4ac857be-79cb-5e3b-b425-2c701c47a2c9 | /documentation/ops-monitoring/hybrid-cloud-datadog/ | Monitoring with Datadog | Prerequisites | documentation | false | Monitoring with Datadog > Prerequisites
- A Kubernetes cluster with a Qdrant database deployed, using either a Hybrid Cloud or Private Cloud deployment
- `kubectl` and `helm` configured to interact with your cluster
- A Datadog account and a [Datadog API key](https://docs.datadoghq.com/account_management/api-app-keys/... | [
0.007625679951161146,
-0.029423601925373077,
-0.06356373429298401,
-0.030932307243347168,
0.05915307253599167,
-0.039104118943214417,
0.07114116102457047,
0.02784709259867668,
0.027380896732211113,
-0.03367135301232338,
0.006325974594801664,
0.03387932479381561,
0.06624328345060349,
-0.033... |
4ad0c96d-0808-5492-a16e-8a1966f1eacc | /documentation/tutorials-build-essentials/video-anomaly-edge-part-2/ | Video Anomaly Detection Part 2: Edge-to-Cloud Pipeline | Edge-to-Cloud Escalation > Ensemble Scoring | documentation | true | Video Anomaly Detection Part 2: Edge-to-Cloud Pipeline > Edge-to-Cloud Escalation > Ensemble Scoring

The ensemble weighting reflects the accuracy differential bet... | [
0.0544540174305439,
-0.006495279725641012,
-0.07330470532178879,
-0.07061264663934708,
0.01594780571758747,
-0.009420368820428848,
0.052439577877521515,
0.04027397558093071,
0.08635297417640686,
-0.02918555960059166,
0.013967979699373245,
0.057269614189863205,
0.03247446566820145,
-0.00465... |
4ad2d933-e0b4-5142-8f17-12dd8f1bcad3 | /documentation/tutorials-develop/code-search/ | Semantic Search for Code | Querying the codebase | documentation | true | Semantic Search for Code > Querying the codebase
We use one of the models to search the collection. Start with text embeddings.
Run the following query "*How do I count points in a collection?*". Review the
results.
In these tables, we link to longer code excerpts from a
`file_name` in the `Qdrant` repository. The re... | [
0.06290671974420547,
-0.05472775176167488,
-0.02493663877248764,
0.0070391325280070305,
0.04205813631415367,
-0.022565685212612152,
0.06813226640224457,
0.07172922790050507,
0.04361332580447197,
-0.005004432518035173,
0.050458066165447235,
0.03771889582276344,
0.03109021857380867,
-0.03475... |
4ad4af7e-ebd5-5b1d-bffb-203003e4e14f | /documentation/ops-optimization/optimize/ | Optimize Performance | documentation | false | Optimize Performance
Different use cases require different balances between memory usage, search speed, and precision. Qdrant is designed to be flexible and customizable so you can tune it to your specific needs.
This guide will walk you three main optimization strategies:
- High Speed Search & Low Memory Usage
- Hi... | [
0.02230099029839039,
-0.05259406566619873,
-0.028207359835505486,
-0.008793028071522713,
0.08893203735351562,
-0.04029804468154907,
0.03992033004760742,
0.005762782879173756,
0.09028460085391998,
-0.042851973325014114,
-0.003297724761068821,
0.0059461151249706745,
0.04515357315540314,
-0.0... | |
4ad6bec0-6dfe-5f3c-9987-759c798f1161 | /documentation/tutorials-search-engineering/query-decomposition/ | Query Decomposition for Multi-Hop Questions | documentation | false | Query Decomposition for Multi-Hop Questions
| Time: 15 min | Level: Intermediate | Stack: Python |
|--------------|---------------------|---------------|
A multi-hop question chains two facts: the second depends on the answer to the first. "Where was the director of the film Inception born?" needs the director, then ... | [
0.04143749922513962,
-0.04834173619747162,
-0.04061749577522278,
-0.03943420946598053,
0.06034097820520401,
-0.02993677370250225,
0.03727327659726143,
0.005263907834887505,
0.08032110333442688,
-0.02426847256720066,
0.044484518468379974,
0.00678184162825346,
0.06905099004507065,
-0.0389127... | |
4ae9e812-fc6e-51d8-a223-86ac538cef72 | /documentation/tutorials-operations/large-scale-search/ | Large-Scale Search | Uploading and Indexing | documentation | true | Large-Scale Search > Uploading and Indexing
We will use the following [python script](https://github.com/qdrant/laion-400m-benchmark/blob/master/upload.py) to upload dataset chunks one by one.
```bash
export QDRANT_URL="https://xxxx-xxxx.xxxx.cloud.qdrant.io"
export QDRANT_API_KEY="xxxx-xxxx-xxxx-xxxx"
python upload... | [
0.025556983426213264,
-0.04330754652619362,
-0.05555570498108864,
-0.0071570659056305885,
0.07485119253396988,
-0.03153619170188904,
0.027579911053180695,
0.030351661145687103,
0.08963315933942795,
-0.005732577294111252,
0.017041806131601334,
0.010418477468192577,
0.0708070695400238,
-0.01... |
4afacda4-bb8c-5779-8193-4f796dfa8ff2 | /course/beginners/module-1/why-similarity-is-not-enough/ | Why Similarity Alone Is Not Enough | course | false | Why Similarity Alone Is Not Enough
Module 1
[What Is Search?](/course/beginners/module-1/what-is-search/) and [Why Keyword Search Struggles](/course/beginners/module-1/why-keyword-search-struggles/) showed keyword search failing on synonyms, paraphrasing, polysemy, and word order. It's tempting to read that as "seman... | [
0.051352132111787796,
-0.05530029162764549,
-0.06289386004209518,
-0.01881704293191433,
0.06500817090272903,
-0.021949611604213715,
0.06927187740802765,
0.01673172600567341,
0.08410463482141495,
-0.03371891751885414,
0.03397272154688835,
0.020426852628588676,
0.06679525226354599,
0.0027388... | |
4b0f013a-d38c-5a20-be19-2ca2a156dccb | /documentation/frameworks/cognee/ | Cognee | How It Works | documentation | false | Cognee > How It Works
Cognee's memory pipelines move content through extraction → embedding → graph construction → retrieval, with consistent configuration across laptops, distributed jobs, and hosted runs. The graph-aware semantic layer is where retrieval becomes reasoning.
Here's what stands out in production:
- *... | [
0.022477632388472557,
-0.0214998759329319,
-0.06490923464298248,
-0.056407004594802856,
0.056803397834300995,
-0.029912203550338745,
0.06026376038789749,
0.015642253682017326,
0.05960458144545555,
-0.013993168249726295,
0.020319441333413124,
-0.00623187143355608,
0.05299049988389015,
-0.03... |
4b0ff5cc-f408-514a-9a49-fe4822b6edde | /documentation/migration-guidance/pre-migration-baseline/ | Pre-Migration Baseline | What to Capture > 2. Metadata Sample | documentation | true | Pre-Migration Baseline > What to Capture > 2. Metadata Sample
Export a representative sample of metadata (or payloads) from your source system. You'll use this for field-by-field comparison after migration.
**How much to sample:** At least 1,000 records or 1% of your data, *whichever is larger*. For datasets under 10... | [
0.04433870688080788,
-0.05836591124534607,
-0.03318765014410019,
-0.04419543221592903,
0.03133402392268181,
-0.03806150704622269,
0.06483785063028336,
0.04898669570684433,
0.06901232153177261,
-0.07253514975309372,
0.03305031731724739,
0.004391844384372234,
0.06036622077226639,
-0.02472918... |
4b225c44-485d-5fbb-bebd-7ce2eb3b3c28 | /documentation/ops-monitoring/monitoring/ | Monitoring & Telemetry | Cluster-Wide Telemetry | documentation | false | Monitoring & Telemetry > Cluster-Wide Telemetry
The `/telemetry` endpoint reports from the point of view of the peer being queried. Qdrant also provides a `/cluster/telemetry` endpoint, which aggregates telemetry from all peers.
This includes less information than `/telemetry`, but provides information like shard tra... | [
0.0287794042378664,
-0.01711859367787838,
-0.0460125133395195,
-0.04262715205550194,
0.02523711323738098,
-0.011208312585949898,
0.0813925489783287,
0.09153228253126144,
0.07197027653455734,
-0.03297167271375656,
0.045534633100032806,
-0.007491229567676783,
0.03756129741668701,
-0.00027837... |
4b543e76-a39a-5f8e-b422-09c7cfd4e0ce | /articles/hyperbolic-embeddings-qdrant/ | Why Hyperbolic Embeddings Matter for Hierarchical Data | Testing Hyperbolic Embeddings | article | false | Why Hyperbolic Embeddings Matter for Hierarchical Data > Testing Hyperbolic Embeddings
WordNet is useful for showing why hyperbolic embeddings work, but we also wanted to see what happens on something closer to a real application. So we tested the same idea on the Google Product Taxonomy: 5,595 categories, seven level... | [
0.052315663546323776,
0.013651609420776367,
-0.04085799679160118,
-0.044730693101882935,
0.05253692343831062,
-0.04572676867246628,
0.04605184495449066,
0.06915126740932465,
0.09725217521190643,
-0.021687351167201996,
0.007440441288053989,
0.014105947688221931,
0.06803740561008453,
-0.0418... |
4b6cea06-add4-55f7-a738-b2986e46c3a3 | /articles/relevance-feedback/ | Relevance Feedback in Qdrant | Feedback-Based Scoring > Training Objective | article | false | Relevance Feedback in Qdrant > Feedback-Based Scoring > Training Objective
With the right a, b and c, adjusted to dataset, retriever and feedback scores distribution, the naive formula should rank documents better than our simple retriever.
That leads us to **minimizing Pairwise Ranking Loss** as the training object... | [
0.02398901991546154,
-0.026329560205340385,
-0.042464908212423325,
-0.039776142686605453,
0.06536246836185455,
-0.014015168882906437,
0.014064677059650421,
0.02862822450697422,
0.07996125519275665,
0.00628673518076539,
0.008659228682518005,
0.028124623000621796,
0.10772458463907242,
-0.026... |
4b84bdba-2a8a-526b-9bf6-0149c9d069e7 | /course/essentials/day-0/qdrant-cloud/ | Qdrant Setup | Create your cluster | course | false | Qdrant Setup > Create your cluster
1. Sign up at [cloud.qdrant.io](https://cloud.qdrant.io/signup) with email, Google, or GitHub.
2. Open **Clusters** → **Create a Free Cluster**. The Free Tier is enough for this course.

3. Pick a region close... | [
0.026557069271802902,
-0.03791207820177078,
-0.02795819193124771,
-0.02025025524199009,
0.05399245396256447,
-0.03678222373127937,
0.048393379896879196,
0.028542062267661095,
0.03714462369680405,
-0.03744204714894295,
0.021982364356517792,
0.0342903733253479,
0.0862739160656929,
-0.0399661... |
4b8f3d07-413b-5a81-a301-403f0b2e09ba | /documentation/private-cloud/private-cloud-setup/ | Setup Private Cloud | Requirements > CLI tools | documentation | false | Setup Private Cloud > Requirements > CLI tools
During the onboarding, you will need to deploy the Qdrant Kubernetes Operator using Helm. Make sure you have the following tools installed:
* [kubectl](https://kubernetes.io/docs/tasks/tools/install-kubectl/)
* [helm](https://helm.sh/docs/intro/install/)
You will need t... | [
0.01909707859158516,
-0.07747763395309448,
-0.014989715069532394,
-0.018284007906913757,
0.061027877032756805,
-0.05143814906477928,
0.02440824545919895,
0.03425711765885353,
0.028704877942800522,
-0.01225313637405634,
0.012395743280649185,
0.018925724551081657,
0.04188062250614166,
-0.055... |
4ba856ab-5bd3-5e27-975f-5da28bf952c8 | /documentation/tutorials-search-engineering/sparse-embeddings-ecommerce-part-4/ | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 4: Specialization vs Generalization | Decision Framework | documentation | false | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 4: Specialization vs Generalization > Decision Framework
After running all these experiments, here's when to use each approach:
| Scenario | Recommended approach |
|---|---|
| Single retailer, lots of training data | **Domain-specific fine-tuning**: maximum p... | [
0.027011428028345108,
-0.021572181954979897,
-0.049887996166944504,
-0.03919906169176102,
0.03871435299515724,
-0.0478869304060936,
0.07075663655996323,
-0.0007677712710574269,
0.04620780423283577,
-0.006128538399934769,
0.012408294714987278,
0.040389109402894974,
0.05999217554926872,
-0.0... |
4baba71c-5855-5d02-b3fc-d26010c5f0fc | /documentation/tutorials-build-essentials/video-anomaly-edge-part-3/ | Video Anomaly Detection Part 3: Scoring, Governance, and Deployment | Results | documentation | false | Video Anomaly Detection Part 3: Scoring, Governance, and Deployment > Results

```python
from qdrant_client import QdrantClient, models
client = QdrantClient(url="http://localhost:6333")
client.query_points(
collec... | [
0.030165638774633408,
-0.08001502603292465,
-0.039124246686697006,
-0.02110668271780014,
0.07411115616559982,
-0.01944912038743496,
0.041049376130104065,
0.004462047945708036,
0.07486988604068756,
-0.01750551536679268,
0.06031037122011185,
0.04008115828037262,
0.03465603291988373,
-0.05355... |
4bd8c695-8bc4-5244-aeec-41541a9fdb5f | /documentation/embeddings/voyage/ | Voyage AI | Reduce Vector Size | documentation | true | Voyage AI > Reduce Vector Size
The models in the Models table with several dimensions accept an `output_dimension` parameter (`outputDimension` in TypeScript). Set the collection's `size` to the same value, and pass the same `output_dimension` when you embed queries:
```python
response = vo.embed(
documents,
... | [
0.06029103323817253,
-0.0031139226630330086,
-0.010977261699736118,
-0.06176789104938507,
0.04837630316615105,
-0.007527715060859919,
0.051109571009874344,
0.033901508897542953,
0.06269627809524536,
-0.0324426144361496,
0.04493628814816475,
0.05383402854204178,
0.07667273283004761,
-0.0131... |
4bf264d8-2643-5fd1-8669-0d635e8bee92 | /documentation/migrate-to-qdrant/from-milvus/ | From Milvus | Gotchas | documentation | false | From Milvus > Gotchas
- **Partitions:** The tool does not record which partition a point came from, so points from all migrated partitions are indistinguishable in the target collection.
## Next Steps
After migration, verify your data arrived correctly with the [Migration Verification Guide](/documentation/migration... | [
0.02624177187681198,
-0.059871286153793335,
-0.021051844581961632,
-0.038016851991415024,
0.050262123346328735,
0.02852059155702591,
0.0330682247877121,
0.039923638105392456,
0.0802098959684372,
-0.007261822000145912,
0.010262358002364635,
0.017895936965942383,
0.03695059195160866,
-0.0315... |
4c0e27ed-0fbb-5995-872a-f3dc9d638578 | /documentation/cloud/authentication/ | Authentication | Authenticate via SDK | documentation | true | Authentication > Authenticate via SDK
Now that you have created your first cluster and key, you might want to access your database from within your application.
Our [official Qdrant clients](/documentation/interfaces/) for Python, TypeScript, Go, Rust, .NET and Java all support the API key parameter.
```bash
curl \
... | [
0.01197836920619011,
-0.024606430903077126,
-0.013300525024533272,
-0.015272813849151134,
0.04343009367585182,
-0.0514054037630558,
0.044845812022686005,
0.040374506264925,
0.05246954783797264,
-0.02212895266711712,
0.04832354187965393,
0.030518393963575363,
0.0844179317355156,
-0.02139188... |
4c1b181f-fa39-529a-810a-d2e08eeb34ee | /documentation/capacity-planning/ | Capacity Planning | Calculating RAM and Disk Size > ID Tracker | documentation | true | Capacity Planning > Calculating RAM and Disk Size > ID Tracker
The ID tracker maps each point's external ID to its internal storage location and current version. Qdrant persists it on disk and, in the default `pinned` tier, also keeps it in RAM:
```text
id_tracker_size = base * 52 bytes
```
For example, with one mil... | [
0.022732684388756752,
-0.0345938615500927,
-0.02077658288180828,
-0.035195957869291306,
0.04970889911055565,
-0.005978154018521309,
0.07031615823507309,
0.04908359795808792,
0.1239379420876503,
-0.018968403339385986,
0.003972657024860382,
-0.00699247233569622,
0.04349256306886673,
0.012467... |
4c45b4a0-025d-5559-b3ab-2003ed680f8e | /documentation/private-cloud/api-reference/ | API Reference | qdrant.io/v1 > Resource Types > QdrantClusterSpec | documentation | false | API Reference > qdrant.io/v1 > Resource Types > QdrantClusterSpec
| Field | Description | Default | Validation |
| --- | --- | --- | --- |
| `onDemandReplication` _[OnDemandReplicationType](#ondemandreplicationtype)_ | OnDemandReplication specifies the on-demand replication restart mode.Off (default): Normal restart b... | [
0.037860848009586334,
-0.034382689744234085,
-0.0363202840089798,
-0.02658972702920437,
0.07445593923330307,
-0.06121120974421501,
0.0425766296684742,
0.051504235714673996,
0.03812473639845848,
-0.0394468680024147,
0.053357113152742386,
0.022785956040024757,
0.026256676763296127,
-0.002467... |
4c81458b-b2ee-53b9-8686-7e9d291a3787 | /articles/late-interaction-models/ | Late Interaction Retrieval with Dense Token Embeddings | You Can Adapt Dense Embedding Models for Late Interaction | article | false | Late Interaction Retrieval with Dense Token Embeddings > You Can Adapt Dense Embedding Models for Late Interaction
Qdrant 1.10 introduced support for multi-vector representations, with late interaction being a prominent example of this model. In essence, both documents and queries are represented by multiple vectors, ... | [
0.05483342334628105,
-0.03710448741912842,
-0.055398739874362946,
-0.05612796172499657,
0.0720224678516388,
-0.0644112080335617,
0.054831307381391525,
0.043380726128816605,
0.07835239917039871,
-0.03143998235464096,
0.04481976106762886,
0.014279017224907875,
0.04531516507267952,
-0.0427612... |
4c88990d-f63d-5330-bcaf-0537d2bc4886 | /articles/vector-similarity-beyond-search/ | Vector Similarity: Going Beyond Full-Text Search | Dissimilarity Search > Detecting Mislabeled Items | article | true | Vector Similarity: Going Beyond Full-Text Search > Dissimilarity Search > Detecting Mislabeled Items
For example, we have a dataset of furniture in which we have classified our items into what kind of furniture they are: tables, chairs, and lamps.
To ensure our catalog is accurate, we can use a dissimilarity search to... | [
0.09151028841733932,
-0.03227291256189346,
-0.047791507095098495,
-0.02852424792945385,
0.07943430542945862,
-0.030711324885487556,
0.09887585043907166,
0.052254535257816315,
0.06300213187932968,
-0.05475557595491409,
0.03858884051442146,
0.06242373213171959,
0.05211462080478668,
-0.009138... |
4ca428cd-1255-56f1-a8dc-70aed48598ae | /articles/io_uring/ | Faster Disk I/O for Vector Search Using io_uring | article | false | Faster Disk I/O for Vector Search Using io_uring
With Qdrant [version 1.3.0](https://github.com/qdrant/qdrant/releases/tag/v1.3.0) we
introduce the alternative io\_uring based *async uring* storage backend on
Linux-based systems. Since its introduction, io\_uring has been known to improve
async throughput wherever the... | [
0.04492245614528656,
-0.062237925827503204,
-0.03707830235362053,
0.00041519655496813357,
0.07885891199111938,
-0.01329112146049738,
0.04860929772257805,
-0.01655937358736992,
0.0756482258439064,
0.010645787231624126,
0.010540205985307693,
0.004614762496203184,
0.05709587782621384,
-0.0138... | |
4cb3977a-341c-5c34-8d24-7154a694f57a | /documentation/examples/rag-chatbot-vultr-dspy-ollama/ | Private RAG Information Extraction Engine | Wrapping up | documentation | false | Private RAG Information Extraction Engine > Wrapping up
In this tutorial, we showed you how to set up a private environment for information extraction using DSPy, Ollama, and
Qdrant. All the components might be securely hosted on the Vultr cloud, giving you full control over your data. | [
0.06703253835439682,
-0.016893276944756508,
-0.04769632965326309,
-0.02180107869207859,
0.08226871490478516,
-0.030176963657140732,
0.07509537786245346,
0.06424394994974136,
0.042400963604450226,
-0.027268482372164726,
0.014971869997680187,
0.031890030950307846,
0.0838932916522026,
-0.0317... |
4cc66f4d-965a-5cb3-820e-d7a556ba9201 | /documentation/platforms/buildship/ | BuildShip | documentation | false | BuildShip
[BuildShip](https://buildship.com/) is a low-code visual builder to create APIs, scheduled jobs, and backend workflows with AI assistance.
You can use the [Qdrant integration](https://buildship.com/integrations/qdrant) to development workflows with semantic-search capabilities.
## Prerequisites
1. A Qdran... | [
0.012701934203505516,
-0.034737780690193176,
-0.08023612201213837,
-0.02463124692440033,
0.0457981713116169,
-0.048389434814453125,
0.06946104764938354,
0.0781000554561615,
0.015357804484665394,
-0.029090994969010353,
0.027340443804860115,
-0.005585641600191593,
0.040147311985492706,
-0.03... | |
4ccc6941-1f38-50cb-935e-819b250e984d | /documentation/cloud-rbac/role-management/ | Role Management | Built-In Roles | documentation | false | Role Management > Built-In Roles
Qdrant Cloud includes some built-in roles for common use-cases. The permissions for these built-in roles cannot be changed.
There are three types:
- The **Base Role** is assigned to all users, and provides the minimum privileges required to access Qdrant Cloud.
- The **Admin Role** ... | [
0.002466502133756876,
-0.03365502506494522,
-0.033754657953977585,
-0.050947632640600204,
0.04846592992544174,
-0.04870898276567459,
0.04570090025663376,
0.05748842656612396,
0.09117443859577179,
-0.035668015480041504,
0.014205695129930973,
0.00923637580126524,
0.06568817049264908,
-0.0412... |
4cce50b2-6194-5fdc-872d-540fe471e3a0 | /documentation/tutorials-search-engineering/ann-recall/ | Measuring ANN Recall | Next Steps | documentation | false | Measuring ANN Recall > Next Steps
Once ANN recall is on target, continue with [Measuring Retrieval Relevance](/documentation/search-evaluation/retrieval-relevance/) to check how well those results match user intent. | [
0.015712086111307144,
-0.04501133784651756,
0.008357306011021137,
-0.053483135998249054,
0.04541933164000511,
-0.009784366935491562,
0.04247858375310898,
0.012136213481426239,
0.08457204699516296,
-0.010942835360765457,
0.025805748999118805,
0.02867707423865795,
0.08196885883808136,
-0.002... |
4cd15ec2-4241-56f8-9376-22a425dd1bdd | /documentation/tutorials-operations/time-based-sharding/ | Time-Based Sharding | Ingest Historical Data | documentation | false | Time-Based Sharding > Ingest Historical Data
- The CSV file is streamed row by row, buffering points in batches of 100 for efficient uploading. The optimal batch size depends on your data and cluster, so you may want to experiment with different sizes for best performance.
- The date (`YYYY-MM-DD`) is extracted from e... | [
0.01256399042904377,
-0.01989533193409443,
-0.061963118612766266,
-0.039244093000888824,
0.054187577217817307,
-0.014405527152121067,
0.022302452474832535,
0.022952796891331673,
0.07952244579792023,
-0.008392981253564358,
0.02716246247291565,
-0.005041392985731363,
0.025502195581793785,
-0... |
4cd497df-8460-516d-a7e4-af87d12f00e6 | /documentation/cloud-api/ | Qdrant Cloud API | Introduction > Samples | documentation | false | Qdrant Cloud API > Introduction > Samples
For samples on how to use the API, with a tool like grpcurl, curl or any of the provided SDKs, please see the [Qdrant Cloud Public API](https://github.com/qdrant/qdrant-cloud-public-api) repository.
## Terraform Provider
Qdrant Cloud also provides a Terraform provider to man... | [
-0.008211598731577396,
-0.016287384554743767,
-0.013090893626213074,
-0.03875291347503662,
0.02358640730381012,
-0.008169742301106453,
0.1174604520201683,
0.05399497598409653,
0.06778556853532791,
-0.050763025879859924,
0.04327006638050079,
-0.0041640871204435825,
0.05870930850505829,
-0.0... |
4cd7737a-dbdb-599a-8333-f3c48e23e727 | /documentation/tutorials-build-essentials/video-anomaly-edge-part-2/ | Video Anomaly Detection Part 2: Edge-to-Cloud Pipeline | Two-Shard Edge Architecture > Representative Subset Selection | documentation | true | Video Anomaly Detection Part 2: Edge-to-Cloud Pipeline > Two-Shard Edge Architecture > Representative Subset Selection
The cloud baseline may contain 100,000+ normal vectors. The edge device only needs 500. We use MiniBatchKMeans to cluster the cloud embeddings and select the vector closest to each centroid:
```pytho... | [
0.01868337020277977,
-0.014563119038939476,
-0.03350382298231125,
-0.0681111067533493,
0.051864542067050934,
-0.04138907790184021,
0.0645168349146843,
0.051373038440942764,
0.06888136267662048,
-0.028033442795276642,
0.06765426695346832,
0.017695002257823944,
0.014247986488044262,
-0.02259... |
4cdcf76a-cbf1-5aea-9f3e-8837b5903ed8 | /documentation/search-evaluation/retrieval-relevance/ | Measuring Retrieval Relevance | Generating Queries > 1. Human Annotation | documentation | false | Measuring Retrieval Relevance > Generating Queries > 1. Human Annotation
Domain experts assign relevance scores on a binary (relevant / not relevant) or graded (0/1/2 or 1–5) scale. Human-labeled data produces the highest-fidelity signal and is the primary source for graded labels. Expert time is the bottleneck, which... | [
0.0030623034108430147,
-0.02736041136085987,
-0.06249899044632912,
-0.05133140832185745,
0.02047034166753292,
-0.04450878128409386,
0.051683954894542694,
0.060749538242816925,
0.07041283696889877,
-0.05067641660571098,
0.042691510170698166,
0.007331056986004114,
0.09268463402986526,
-0.013... |
4ce2136c-32f3-5171-9de9-d1528f2ba132 | /documentation/tutorials-operations/prevent-unoptimized-usage/ | Prevent Unoptimized Usage | Create Two Collections | documentation | true | Prevent Unoptimized Usage > Create Two Collections
Create two collections of 768-dimensional vectors, one with `prevent_unoptimized` enabled and one without, to compare query latency and optimization time between them.
```python
from qdrant_client import models
async def create_collection(collection_name: str, preve... | [
0.032984934747219086,
-0.04001783952116966,
-0.05862479656934738,
-0.041103776544332504,
0.06907164305448532,
-0.04885810613632202,
0.051586706191301346,
0.026723213493824005,
0.09799083322286606,
-0.04623091593384743,
0.04677409306168556,
0.02779405005276203,
0.09183415025472641,
-0.01236... |
4cf8c8fa-b45b-5d2d-ad66-3dbdde8d0b46 | /articles/minicoil/ | miniCOIL: on the Road to Usable Sparse Neural Retrieval | Key Takeaways > The Right Tool for the Right Job | article | false | miniCOIL: on the Road to Usable Sparse Neural Retrieval > Key Takeaways > The Right Tool for the Right Job
When are miniCOIL retrievers applicable?
If you need precise term matching but BM25-based retrieval doesn't meet your needs, ranking higher documents with words of the right form but the wrong semantical meaning... | [
0.012058538384735584,
-0.04268724098801613,
-0.028874428942799568,
-0.04899638891220093,
0.04583640769124031,
-0.04994671046733856,
0.05716981366276741,
0.04702308773994446,
0.07261981815099716,
-0.015301009640097618,
0.023210519924759865,
0.019389232620596886,
0.06538628041744232,
-0.0125... |
4d102d4d-0ae6-5c5a-b4cb-a775764c9862 | /documentation/cloud/cluster-monitoring/ | Monitor Clusters | Alerts | documentation | false | Monitor Clusters > Alerts
Hotspotting, where a subset of nodes handle more load than the rest, causes slow search performance and potentially node failures.*
For uneven CPU usage it typically means data distribution across nodes is uneven.
**What can I do to resolve this?**
Resharding and shard rebalancing are the... | [
0.01874246448278427,
-0.030606234446167946,
-0.0640544593334198,
-0.049353182315826416,
0.06584474444389343,
-0.04033400118350983,
0.03192770853638649,
0.032713763415813446,
0.08998319506645203,
-0.053842246532440186,
0.022034820169210434,
-0.01173386536538601,
0.03379352390766144,
-0.0060... |
4d286c03-61a9-5601-b5ff-fcd95dcacce0 | /course/multi-vector-search/module-2/how-colpali-works/ | How ColPali Models Work | From Text to Visual Documents | course | false | How ColPali Models Work > From Text to Visual Documents
**What about documents that aren't just text?** PDFs often contain diagrams, tables, charts, equations, and complex layouts where the visual presentation carries as much meaning as the text itself.
Traditional approaches to searching visual documents typically i... | [
0.0246480293571949,
-0.05692075565457344,
-0.0682755708694458,
-0.05904901400208473,
0.07672233134508133,
0.003800350008532405,
0.022626372054219246,
0.04405239596962929,
0.056875597685575485,
0.0018747479189187288,
0.03341514989733696,
-0.01724846474826336,
0.05366897955536842,
-0.0082421... |
4d2e9c28-7c7d-5476-9a2a-4b6715e10e29 | /course/essentials/day-0/building-simple-vector-search/ | Implementing a Basic Vector Search | Step 1: Install the Qdrant Client | course | true | Implementing a Basic Vector Search > Step 1: Install the Qdrant Client
To interact with Qdrant, we need the Python client. This enables us to communicate with the Qdrant service, manage collections, and perform vector searches.
```python
!pip install qdrant-client
``` | [
-0.0014387810369953513,
-0.0856543779373169,
-0.039739642292261124,
0.0032041871454566717,
0.05541989207267761,
-0.048059433698654175,
0.05718290060758591,
0.035150494426488876,
0.060465291142463684,
-0.018633665516972542,
0.028564434498548508,
0.03480791673064232,
0.05333331972360611,
-0.... |
4d3d7403-bdfa-57b4-8222-28142de119ed | /documentation/tutorials-build-essentials/rag-deepseek/ | 5-Minute RAG with DeepSeek | documentation | false | 5-Minute RAG with DeepSeek
| Time: 5 min | Level: Beginner |
| --- | --- |
This tutorial demonstrates how to build a **Retrieval-Augmented Generation (RAG)** pipeline using FastEmbed for embeddings, Qdrant for vector search, and DeepSeek to answer questions with retrieved context. RAG pipelines enhance Large Language... | [
0.027391275390982628,
-0.039102837443351746,
-0.030092984437942505,
-0.03965967521071434,
0.07823097705841064,
-0.04143266752362251,
0.052784282714128494,
0.041870519518852234,
0.04937286674976349,
-0.048832837492227554,
0.01857088692486286,
0.02666546404361725,
0.04746824875473976,
-0.036... | |
4d3e8886-7617-561c-b23c-4de3a694d917 | /documentation/tutorials-operations/embedding-model-migration/ | Migrate to a New Embedding Model | Blue-Green Migration > Step 4: Change the Collection and Embedding Model for Searches | documentation | true | Migrate to a New Embedding Model > Blue-Green Migration > Step 4: Change the Collection and Embedding Model for Searches
Once the migration process is complete and all the points from the old collection are re-embedded and stored in the new collection, [compare retrieval quality](#check-retrieval-quality-before-switch... | [
0.03974548727273941,
-0.06715002655982971,
-0.03167496994137764,
-0.022295504808425903,
0.04332229867577553,
-0.017955735325813293,
0.025332853198051453,
0.042702727019786835,
0.06037743762135506,
-0.029010076075792313,
0.03530444949865341,
0.0389314740896225,
0.029136382043361664,
-0.0405... |
4d557370-2cde-517e-980a-02317461a32d | /documentation/tutorials-search-engineering/sparse-embeddings-ecommerce-part-4/ | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 4: Specialization vs Generalization | Series Summary | documentation | false | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 4: Specialization vs Generalization > Series Summary
- **[Part 1: Why Sparse Embeddings Beat BM25](/documentation/tutorials-search-engineering/sparse-embeddings-ecommerce-part-1/)** - SPLADE combines keyword precision with learned expansion
- **[Part 2: Traini... | [
0.04703386873006821,
-0.02648847922682762,
-0.048827558755874634,
-0.03671140968799591,
0.04790946841239929,
-0.06953137367963791,
0.06646255403757095,
0.036799293011426926,
0.059668947011232376,
-0.00336679769679904,
0.014181115664541721,
0.020103735849261284,
0.030189868062734604,
-0.019... |
4d6519bf-d0dc-5d95-9cf1-fc7a5cbf0224 | /course/essentials/day-1/embedding-models/ | Points, Vectors and Payloads | Vector Types in Qdrant > Sparse Vectors | course | true | Points, Vectors and Payloads > Vector Types in Qdrant > Sparse Vectors
Mathematically identical to dense vectors, but containing many zeros. They use optimized storage representation and have a different shape than dense vectors.
**Representation:**
Sparse vectors are represented as a list of (index, value) pairs:
- ... | [
0.03934537246823311,
-0.035431332886219025,
-0.03597790375351906,
-0.04104487597942352,
0.051648326218128204,
-0.04778197780251503,
0.024951180443167686,
0.031383443623781204,
0.08233880251646042,
-0.03939030319452286,
0.012625142931938171,
0.005451252683997154,
0.02986583299934864,
0.0061... |
4d662039-2a7c-5a7e-bf24-d03e1688cf7a | /articles/minicoil-en-es/ | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier | From Words to Concepts > Standing on the Shoulders of BM25, Again (We Got Comfortable) | article | false | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier > From Words to Concepts > Standing on the Shoulders of BM25, Again (We Got Comfortable)
*At this point, we're jumping on these shoulders with the comfort of a clingy toddler.*
The scoring formula between sparse miniCOIL EN-ES vectors copy-cats the m... | [
0.05610072240233421,
-0.024532925337553024,
-0.025151129812002182,
-0.06568162888288498,
0.06180531531572342,
-0.03708280250430107,
0.07344387471675873,
0.04581889137625694,
0.08001717180013657,
-0.016268441453576088,
0.01585691049695015,
0.01276319194585085,
0.039575111120939255,
-0.03517... |
4d67dd63-3c6c-53c0-9edc-4ae6e15fcbf0 | /documentation/migration-recovery-options/ | Migration and Recovery | Qdrant Migration Tool | documentation | false | Migration and Recovery > Qdrant Migration Tool
The [Qdrant Migration Tool](/documentation/migrate-to-qdrant/) streams points from a source database directly to a target Qdrant cluster over the network. Because data arrives as individual upserts, Qdrant builds the HNSW index from scratch. This is the slowest option for... | [
0.05860743299126625,
-0.07256044447422028,
-0.03456588834524155,
-0.0251576267182827,
0.034607499837875366,
-0.02598665840923786,
0.09609419107437134,
0.03890317305922508,
0.06160486862063408,
-0.014311261475086212,
0.015093543566763401,
0.0007874827715568244,
0.03682311624288559,
-0.01167... |
4d752185-be03-5832-92da-35fb5d5805cd | /documentation/ops-optimization/optimize/ | Optimize Performance | 3. High Precision with High-Speed Search | documentation | true | Optimize Performance > 3. High Precision with High-Speed Search
For scenarios requiring both high speed and high precision, keep as much data in RAM as possible. Apply quantization with re-scoring for tunable accuracy.
Here is how you can configure TurboQuant quantization for a collection:
```python
from qdrant_clie... | [
0.0013606579741463065,
-0.07221965491771698,
-0.0388699546456337,
-0.031667739152908325,
0.039014510810375214,
-0.014565517194569111,
0.06319227814674377,
-0.016962919384241104,
0.06998219341039658,
-0.02834412083029747,
0.013985431753098965,
0.041383616626262665,
0.061933670192956924,
-0.... |
4d84b163-28b6-519a-b283-d202248e3619 | /documentation/scaling/node-failure-recovery/ | Node Failure Recovery | Recover from a Snapshot | documentation | false | Node Failure Recovery > Recover from a Snapshot
If shards are unrecoverable and there are no other copies in the cluster, you can still recover from a snapshot.
First, detach the failed node and create a new one:
1. To exclude failed nodes from the consensus, use [remove peer](https://api.qdrant.tech/master/api-refe... | [
0.041589412838220596,
-0.019154049456119537,
-0.02736000530421734,
-0.04356417804956436,
0.05313422158360481,
-0.009387912228703499,
0.030642826110124588,
0.013019510544836521,
0.04649282246828079,
-0.05612647905945778,
0.04623367264866829,
0.05028147250413895,
0.037386320531368256,
-0.023... |
4db9ca8a-0eba-5bcb-8d1a-b3f175128884 | /documentation/frameworks/cognee/ | Cognee | Why Qdrant For The Memory Layer | documentation | false | Cognee > Why Qdrant For The Memory Layer
At runtime, [Cognee](https://www.cognee.ai/)'s semantic memory layer requires fast and predictable lookups to surface candidates for graph reasoning, as well as tight control over metadata to ground multi-hop traversals. Qdrant's design aligns with those needs with its:
- Near... | [
0.023701127618551254,
-0.047162726521492004,
-0.03708064556121826,
-0.04651767760515213,
0.07180391997098923,
-0.018114550039172173,
0.0744963213801384,
0.03654780611395836,
0.06435713917016983,
-0.025691566988825798,
0.013550718314945698,
0.017623284831643105,
0.08005263656377792,
0.00587... |
4dd491d5-c708-5b8b-8fad-7187f1b750b8 | /documentation/tutorials-operations/embedding-model-migration/ | Migrate to a New Embedding Model | Blue-Green Migration > Step 3: Migrate the Existing Points into the New Collection | documentation | true | Migrate to a New Embedding Model > Blue-Green Migration > Step 3: Migrate the Existing Points into the New Collection
Now that you're in dual-write mode, it is time to migrate the existing points from the old collection to the new one. This can be done in a separate process that runs
in parallel with the regular upser... | [
0.04399207979440689,
-0.04331277683377266,
-0.036735352128744125,
-0.017333580181002617,
0.0572931282222271,
-0.0193858090788126,
0.04326196387410164,
0.03930482640862465,
0.05026155337691307,
-0.026603730395436287,
-0.005703279282897711,
0.018320731818675995,
0.015558196231722832,
-0.0291... |
4ddc8092-5edf-51d0-984d-d21a41509d0a | /course/essentials/day-1/embedding-models/ | Points, Vectors and Payloads | Vector Dimensionality | course | false | Points, Vectors and Payloads > Vector Dimensionality
Dense vectors are the most common type used in semantic search and machine learning applications. Vector dimensionality directly impacts search efficiency, memory consumption, and retrieval accuracy.
Higher dimensions capture more detail but cost more in storage an... | [
0.033183880150318146,
-0.01988261379301548,
-0.03145486116409302,
-0.06395137310028076,
0.02009657397866249,
-0.0484941191971302,
0.04389556124806404,
0.07571757584810257,
0.0683959424495697,
-0.05220751836895943,
0.030391545966267586,
0.016025282442569733,
0.03869540989398956,
-0.00996649... |
4dee9d7b-aada-5435-8d08-a6578ac4a920 | /documentation/tutorials-search-engineering/slicing-filter/ | Deterministic Collection Slicing | Setup | documentation | true | Deterministic Collection Slicing > Setup
The tutorial uses top-level awaits as it is adapted from this notebook: if you are using the snippets in a python script, don't forget to use asyncio (or another async runtime like trio) to run the code.
We first need to install the [Qdrant Python Client](https://github.com/qd... | [
0.05632440373301506,
-0.045018184930086136,
-0.04173540696501732,
-0.04016036540269852,
0.045768432319164276,
-0.0363924615085125,
0.10724201798439026,
0.040263887494802475,
0.07433399558067322,
-0.042726486921310425,
0.00566259678453207,
0.04204931855201721,
0.04122158885002136,
-0.024201... |
4df5df77-0251-5472-a364-9f17c3998868 | /documentation/search-precision/reranking-semantic-search/ | Reranking for Better Search | Understanding Reranking > Background | documentation | false | Reranking for Better Search > Understanding Reranking > Background
In search systems, two metrics—precision and recall—are the backbone of success. But what do they mean? Precision tells us how many of the retrieved results are actually relevant, while recall measures how well we’ve captured all the relevant results o... | [
-0.005005002953112125,
-0.062471263110637665,
-0.03651304170489311,
-0.042490970343351364,
0.04341105371713638,
-0.05776330828666687,
0.061148181557655334,
0.04409381002187729,
0.10067960619926453,
-0.031152330338954926,
0.001010758220218122,
0.06359334290027618,
0.03737209737300873,
0.002... |
4e17a676-ff23-5ea9-bd10-bf4c780d1dc2 | /documentation/private-cloud/api-reference/ | API Reference | qdrant.io/v1 > Resource Types > NodeResourceInfo | documentation | false | API Reference > qdrant.io/v1 > Resource Types > NodeResourceInfo
_Appears in:_
- [NodeInfo](#nodeinfo)
| Field | Description | Default | Validation |
| --- | --- | --- | --- |
| `cpu` _string_ | CPU specifies the CPU resources of the node | | |
| `memory` _string_ | Memory specifies the memory resources of the node... | [
0.04421063885092735,
-0.05219174921512604,
0.001415147795341909,
-0.06587115675210953,
0.05901271477341652,
-0.04522901773452759,
0.060375064611434937,
0.07123152911663055,
0.07297584414482117,
-0.06077791377902031,
0.035650625824928284,
-0.03021414764225483,
0.045829884707927704,
-0.01320... |
4e3035ab-1a14-5a91-a714-cf236a129a3d | /articles/dedicated-vector-search/ | Why Vector Search Needs a Dedicated Database | Objections to a Dedicated Service > "A dedicated service means duplicating data" | article | false | Why Vector Search Needs a Dedicated Database > Objections to a Dedicated Service > "A dedicated service means duplicating data"
Embeddings are derived from source data, as described in [Vectors are a Transformation](#vectors-are-a-transformation). Every time you change the embedding model, every vector changes. So eve... | [
-0.00036887655733153224,
-0.0379319004714489,
-0.06168125197291374,
-0.03906184062361717,
0.048147011548280716,
0.0006671066512353718,
0.07567491382360458,
0.05917420983314514,
0.056897349655628204,
-0.012725508771836758,
0.01315799169242382,
0.03135344013571739,
0.0362350158393383,
-0.035... |
4e44eee7-b4b1-5202-9dcf-4315f09724ba | /documentation/tutorials-build-essentials/video-anomaly-edge-part-2/ | Video Anomaly Detection Part 2: Edge-to-Cloud Pipeline | What's Next | documentation | false | Video Anomaly Detection Part 2: Edge-to-Cloud Pipeline > What's Next
In **[Part 3 | Scoring, Governance, and Deployment](/documentation/tutorials-build-essentials/video-anomaly-edge-part-3/)**, we'll cover incident formation from raw scores, baseline governance to prevent poisoning, unified retrieval across cameras, e... | [
0.05958537757396698,
-0.03407823666930199,
-0.0651618018746376,
-0.05004516988992691,
0.06217210739850998,
0.00908934511244297,
0.05026828497648239,
0.012775319628417492,
0.07544787973165512,
-0.034067302942276,
0.07337483763694763,
0.019250329583883286,
0.06666184961795807,
-0.02480039745... |
4e51be9a-c9dc-5f35-8107-b49a175285db | /articles/minicoil-en-es/ | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier | Training miniCOIL EN-ES > miniCOIL EN-ES Backbone | article | false | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier > Training miniCOIL EN-ES > miniCOIL EN-ES Backbone
For miniCOIL EN-ES we need a multilingual input encoder that already "knows" English and Spanish, so that the trainable head focuses on the *sense*, not only on figuring out translation from scratch.... | [
0.044923096895217896,
-0.04145199805498123,
-0.017006391659379005,
-0.03382068872451782,
0.06382708251476288,
-0.07026292383670807,
0.054668113589286804,
0.0466458834707737,
0.0740569531917572,
-0.04931202903389931,
0.006419063080102205,
0.032591670751571655,
0.051429297775030136,
-0.02131... |
4e5b92b4-e964-502a-a6aa-5d92a4e1937f | /course/essentials/day-5/colbert-multivectors/ | Multivectors for Late Interaction Models | ColPali for Visual Documents | course | false | Multivectors for Late Interaction Models > ColPali for Visual Documents
For documents with rich layouts such as PDFs, invoices, and slide decks, ColPali (Contextualized Late Interaction over PaliGemma) extends the same idea to vision. ColPali divides each page into a 32x32 grid (1,024 patches), encodes each patch with... | [
0.02555212937295437,
-0.022813374176621437,
-0.037575092166662216,
-0.05506538972258568,
0.07310225069522858,
-0.029763387516140938,
0.06688753515481949,
0.07098467648029327,
0.014810668304562569,
-0.018504302948713303,
0.010481045581400394,
0.008983063511550426,
0.054306723177433014,
-0.0... |
4e6e74e5-9fba-5b61-8617-f0eda92e06f9 | /articles/minicoil-en-es/ | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier | Training miniCOIL EN-ES | article | false | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier > Training miniCOIL EN-ES
The [v1 recipe](https://qdrant.tech/articles/minicoil/) was: take contextual word representations from an English-only dense encoder, and learn a one-layer projection into a tiny word meaning space, calibrated on how a bigger... | [
0.035658832639455795,
-0.028406919911503792,
-0.021090274676680565,
-0.058862559497356415,
0.07503149658441544,
-0.03529722988605499,
0.03897009417414665,
0.069493867456913,
0.06432218849658966,
-0.016070162877440453,
0.004383429419249296,
0.035880520939826965,
0.08736646920442581,
-0.0155... |
4e76b31f-043d-5b72-9226-b8773db77356 | /documentation/edge/edge-api/ | Edge API | Reference | documentation | false | Edge API > Reference
| Page | What it covers |
| --- | --- |
| [Shard Lifecycle](/documentation/edge/edge-api/shard-lifecycle/) | Creating, loading, inspecting, flushing, and closing an Edge Shard, and reading its metadata with `info` |
| [Configuration](/documentation/edge/edge-api/configuration/) | `EdgeConfig`, den... | [
0.003904187586158514,
-0.01138678565621376,
-0.07331287860870361,
-0.05152256414294243,
0.08567895740270615,
-0.03996695578098297,
0.037501297891139984,
0.05109580606222153,
0.05325033515691757,
-0.03764956817030907,
0.04995448514819145,
0.032638370990753174,
0.035054899752140045,
-0.02069... |
4ea1093d-2382-5d3f-8e81-65620bd36626 | /documentation/private-cloud/api-reference/ | API Reference | qdrant.io/v1 > Resource Types > VolumeSnapshotClass | documentation | false | API Reference > qdrant.io/v1 > Resource Types > VolumeSnapshotClass
_Appears in:_
- [QdrantCloudRegionStatus](#qdrantcloudregionstatus)
| Field | Description | Default | Validation |
| --- | --- | --- | --- |
| `name` _string_ | Name specifies the name of the volume snapshot class | | |
| `driver` _string_ | Driver... | [
0.025033926591277122,
-0.04820605367422104,
0.018857834860682487,
-0.0658048763871193,
0.043740347027778625,
-0.02357029914855957,
0.04563348367810249,
0.04866886883974075,
0.03587816655635834,
-0.04752648249268532,
0.050672318786382675,
0.02284083515405655,
0.05700301751494408,
-0.0248103... |
4eab4da5-26a7-51b6-9f3b-4740b5750682 | /documentation/examples/qdrant-dspy-medicalbot/ | Building a Chain-of-Thought Medical Chatbot with Qdrant and DSPy | Tech Stack Overview | documentation | false | Building a Chain-of-Thought Medical Chatbot with Qdrant and DSPy > Tech Stack Overview
To build a robust and trustworthy medical chatbot, we will combine the following technologies:
* [**Qdrant Cloud**](https://qdrant.tech/cloud/): Qdrant is a high-performance vector search engine for storing and retrieving large col... | [
0.03645588457584381,
-0.03019809164106846,
-0.036754947155714035,
-0.036759328097105026,
0.05664254352450371,
-0.03456924855709076,
0.1046159490942955,
0.026049179956316948,
0.04896680638194084,
-0.03522849082946777,
0.049956340342760086,
-0.0070569999516010284,
0.04112736135721207,
-0.053... |
4eb6ba49-50aa-5aeb-9f86-fd27adef1130 | /documentation/tutorials-build-essentials/video-anomaly-edge-part-1/ | Video Anomaly Detection Part 1: Architecture, Twelve Labs, and NVIDIA VSS | Prerequisites | documentation | false | Video Anomaly Detection Part 1: Architecture, Twelve Labs, and NVIDIA VSS > Prerequisites
**Software**
- **Python 3.12+**: [Download Python](https://python.org/downloads/)
- **Node.js 20+**: [Download Node.js](https://nodejs.org/)
- **uv** (Python package manager): [Install uv](https://docs.astral.sh/uv/)
- **Docker*... | [
0.0368993766605854,
-0.05501090735197067,
-0.07196462899446487,
-0.02288343571126461,
0.06444758176803589,
-0.02181077189743519,
0.05343694984912872,
0.0046890066005289555,
0.08463343977928162,
-0.03126320615410805,
0.046345192939043045,
0.0020949693862348795,
0.07714269310235977,
-0.05087... |
4ecd4c34-c029-5739-b9e9-657da9aac291 | /articles/what-is-vector-quantization/ | What is Vector Quantization? | Rescoring, Oversampling, and Reranking > 2. Oversampling | article | false | What is Vector Quantization? > Rescoring, Oversampling, and Reranking > 2. Oversampling
Oversampling is a technique that helps compensate for any precision lost due to quantization. Since quantization simplifies vectors, some relevant matches could be missed in the initial search. To avoid this, you can **retrieve mor... | [
0.04758721962571144,
-0.0798284113407135,
-0.046358007937669754,
-0.05274046212434769,
0.0554349347949028,
0.016521712765097618,
0.06675691902637482,
-0.009018170647323132,
0.11221853643655777,
-0.025479191914200783,
0.02969701960682869,
0.0015207941178232431,
0.06247854232788086,
-0.01352... |
4ee6bf4e-f9f4-5340-a874-b01660cbeb3e | /course/essentials/day-0/qdrant-cloud/ | Qdrant Setup | Common issues | course | false | Qdrant Setup > Common issues
* **Authentication error**: Recheck the API key and the `api-key` header.
* **Connection error**: Confirm cluster status and region URL; some corporate proxies block outbound TLS. | [
0.04920881986618042,
-0.034835413098335266,
-0.03445510193705559,
-0.029029930010437965,
0.0433499850332737,
-0.03761548176407814,
0.0306374654173851,
0.030318904668092728,
0.06305225193500519,
-0.026852410286664963,
0.009850069880485535,
0.0012742312392219901,
0.06170082464814186,
-0.0437... |
4eea999f-5a2e-5009-943a-add154dbb143 | /course/essentials/day-3/hybrid-search-demo/ | Demo: Implementing a Hybrid Search System | Step 3: Upload the Cheese Dataset | course | true | Demo: Implementing a Hybrid Search System > Step 3: Upload the Cheese Dataset
We're using a small dataset of 10 documents describing different types of cheese and cheese-based dishes. This simple dataset makes it easy to observe the behavior of different search methods:
```python
documents = [
"Aged Gouda develop... | [
0.04670285806059837,
-0.03559587895870209,
-0.06624508649110794,
-0.05430568754673004,
0.08498700708150864,
-0.029794488102197647,
0.05190414935350418,
0.03635120391845703,
0.0601206049323082,
-0.00684414803981781,
0.03579623997211456,
0.007635094225406647,
0.0782485231757164,
-0.038874853... |
4f2f5938-b410-5da9-af80-abd7b543bcf5 | /documentation/data-management/spark/ | Apache Spark | Installation | documentation | true | Apache Spark > Installation
To integrate the connector into your Spark environment, get the JAR file from one of the sources listed below.
- GitHub Releases
The packaged `jar` file with all the required dependencies can be found [here](https://github.com/qdrant/qdrant-spark/releases).
- Building from Source
To bui... | [
0.007154211401939392,
-0.04318808764219284,
-0.059372030198574066,
-0.019727544859051704,
0.05520850792527199,
-0.014441365376114845,
0.02289225347340107,
0.07493472099304199,
0.005278668366372585,
-0.0016907411627471447,
0.022397363558411598,
0.038934722542762756,
0.010272830724716187,
-0... |
4f3d1496-bb89-512c-a6b4-c0082b6c7aeb | /documentation/frameworks/sycamore/ | Sycamore | Reading from Qdrant | documentation | true | Sycamore > Reading from Qdrant
To read a Docset from a Qdrant collection in Sycamore, use the `docset.read.qdrant(....)` function. The Qdrant reader accepts the following arguments:
- `client_params`: Parameters that are passed to the Qdrant client constructor. See more information in the[Client API Reference](https:... | [
0.03888244181871414,
-0.04841247946023941,
-0.023904919624328613,
-0.02957945317029953,
0.06055184826254845,
-0.04616618901491165,
0.06428942829370499,
0.056282881647348404,
0.06864070892333984,
-0.0035603814758360386,
0.04501791298389435,
0.0436837263405323,
0.032469481229782104,
-0.01121... |
4f5d21ab-63a6-5d74-8c26-eb4aded3a283 | /articles/minicoil/ | miniCOIL: on the Road to Usable Sparse Neural Retrieval | Results > Validation Loss | article | false | miniCOIL: on the Road to Usable Sparse Neural Retrieval > Results > Validation Loss
Input transformer [`jina-embeddings-v2-small-en`](https://huggingface.co/jinaai/jina-embeddings-v2-small-en) approximates the “role model” transformer [`mxbai-embed-large-v1`](https://huggingface.co/mixedbread-ai/mxbai-embed-large-v1) ... | [
0.0296087134629488,
-0.04306839406490326,
-0.014098821207880974,
-0.02670000120997429,
0.03362039104104042,
-0.014410015195608139,
0.06699689477682114,
0.06604567170143127,
0.06033613905310631,
-0.04006684571504593,
0.02365792728960514,
0.004540558438748121,
0.028413426131010056,
-0.024018... |
4f7435e9-4bad-5ae9-8e67-77f90d1be170 | /documentation/capacity-planning/ | Capacity Planning | Putting It Together | documentation | true | Capacity Planning > Putting It Together
Add up the components that apply to your setup, based on the tiers you've chosen:
- **RAM**: anything in the `pinned` or `cached` tier:
- Dense vectors (`cached` by default)
- The HNSW vector indexes (`cached` by default)
- Quantized vectors and the sparse indexes (`pinne... | [
0.031911514699459076,
-0.049697309732437134,
-0.04989209771156311,
-0.022787004709243774,
0.06764497607946396,
-0.021085072308778763,
0.06022528186440468,
0.017361167818307877,
0.09744396805763245,
-0.016934609040617943,
0.01784813217818737,
0.008491993881762028,
0.02187364548444748,
-0.02... |
4f9e9362-afac-5d32-97f4-dd8cf73f4820 | /documentation/private-cloud/changelog/ | Changelog | 1.6.4 (2025-04-17) | documentation | false | Changelog > 1.6.4 (2025-04-17)
| Component | Version |
|------------------------|---------|
| qdrant-kubernetes-api | v1.15.5 |
| operator | 2.3.4 |
| qdrant-cluster-manager | v0.3.4 |
* Fix bug in operator Helm chart that caused role binding generation to fail when using `watch.namespa... | [
0.06420779973268509,
-0.03793509304523468,
-0.04765966534614563,
-0.015193387866020203,
0.0697326511144638,
-0.05523919314146042,
0.0385783314704895,
0.06291293352842331,
0.028751306235790253,
-0.024923978373408318,
0.03931370750069618,
0.039862316101789474,
0.04322290048003197,
-0.0443130... |
4fac4cdc-d353-566e-a5c3-0e7126bc8edc | /course/essentials/day-1/movie-search-system/ | Demo: Semantic Movie Search | Step 1: Understanding the Challenge | course | true | Demo: Semantic Movie Search > Step 1: Understanding the Challenge
Our dataset consists of 13 science fiction movies with detailed, literary descriptions. Here's the challenge: each description contains 240-460 tokens, but our embedding model (all-MiniLM-L6-v2) can only embed 256 tokens or less.
**This is where chunki... | [
0.04594952240586281,
-0.013420027680695057,
-0.0857667401432991,
-0.05894685909152031,
0.0853733941912651,
-0.04929294437170029,
-0.014477308839559555,
0.04812540486454964,
0.07021838426589966,
-0.02992074377834797,
0.04362702742218971,
-0.007919601164758205,
0.06431993097066879,
-0.074366... |
4fafceb3-6742-5117-baab-706e38953be4 | /documentation/migration-recovery-options/ | Migration and Recovery | Cloud Backups | documentation | false | Migration and Recovery > Cloud Backups
Cloud Backups work at the disk level: the storage volumes your Qdrant cluster runs on are snapshotted, without Qdrant reading or transferring any data. This makes it the fastest option.
**When to use it:** Recovering from a full cluster failure, or reverting a cluster to a known... | [
0.016582392156124115,
-0.024527287110686302,
-0.006448206026107073,
-0.04972905293107033,
0.06332067400217056,
-0.004189228173345327,
0.0590767040848732,
-0.004796507768332958,
0.06614450365304947,
-0.01795807108283043,
0.018109524622559547,
0.01688937470316887,
0.03496234118938446,
-0.019... |
4fbb0380-7330-5e01-9f58-f420029ac772 | /documentation/examples/hybrid-search-llamaindex-jinaai/ | Chat With Product PDF Manuals Using Hybrid Search | Download the Manuals | documentation | true | Chat With Product PDF Manuals Using Hybrid Search > Download the Manuals
This cell downloads the three PDFs from the [original example's data folder](https://github.com/infoslack/qdrant-example/tree/1a45740d328bae4468dae61e47421eae5024a826/HC-demo/data).
The revision and checksums are fixed so reruns use the same inpu... | [
0.02266712486743927,
-0.018234124407172203,
-0.04404611885547638,
-0.0899808257818222,
0.07235987484455109,
-0.0013911108253523707,
0.04931624233722687,
0.06846074759960175,
0.060677994042634964,
-0.03420890495181084,
-0.014100799337029457,
0.002404424361884594,
0.06906818598508835,
0.0015... |
4fbd268f-e057-5fa4-a893-3710d9d551aa | /articles/predicting-weak-retrieval/ | Predicting Weak Retrieval Without an LLM | Which Signals Fire Depends on the Failure Mode | article | false | Predicting Weak Retrieval Without an LLM > Which Signals Fire Depends on the Failure Mode
Score each signal by how well it separates good from weak retrievals on a labeled split, using AUC (0.5 chance, 1.0 perfect). We ran all of them on three corpora that fail differently: [MuSiQue](https://github.com/StonyBrookNLP/m... | [
0.04978862404823303,
-0.03310871869325638,
-0.07775759696960449,
-0.03765597194433212,
0.07878416031599045,
-0.013797126710414886,
0.034735795110464096,
0.009025008417665958,
0.06073079630732536,
-0.03434545174241066,
0.0038132471963763237,
0.011623378843069077,
0.03183166682720184,
-0.031... |
4fc5674d-5386-5f35-b530-9d7cae315ffd | /documentation/search/search/ | Search | Search API > ACORN Search Algorithm | documentation | true | Search > Search API > ACORN Search Algorithm
*Available as of v1.16.0*
For filtered vector search, you are recommended to create a [payload index](/documentation/manage-data/indexing/#payload-index) for the fields you want to filter by.
During the search, Qdrant will use a combined [filterable index](/documentation/m... | [
0.047814976423978806,
-0.0923321321606636,
-0.05117300525307655,
-0.006898501887917519,
0.0867011621594429,
0.006865629926323891,
0.041983090341091156,
0.0176142118871212,
0.06742967665195465,
-0.0038741272874176502,
0.05932985618710518,
0.02057330496609211,
0.04446778818964958,
-0.0173301... |
500eb170-51a7-5db1-925e-e7b34c139c89 | /documentation/manage-data/vectors/ | Vectors | Vector Types > Multivectors | documentation | true | Vectors > Vector Types > Multivectors
**Available as of v1.10.0**
Qdrant supports the storing of a variable amount of same-shaped dense vectors in a single point.
This means that instead of a single dense vector, you can upload a matrix of dense vectors.
The length of the matrix is fixed, but the number of vectors ... | [
0.04532008618116379,
-0.03343512490391731,
-0.03765934705734253,
-0.03408272936940193,
0.06401679664850235,
-0.032708797603845596,
0.058639269322156906,
0.010766525752842426,
0.09650202840566635,
-0.033546481281518936,
0.054432787001132965,
0.015236769802868366,
0.07704155147075653,
-0.027... |
502a8b1e-c124-5594-836c-d667d7585791 | /documentation/tutorials-search-engineering/sparse-embeddings-ecommerce-part-4/ | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 4: Specialization vs Generalization | Multi-Domain Training > Setting Up Multi-Domain Training | documentation | true | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 4: Specialization vs Generalization > Multi-Domain Training > Setting Up Multi-Domain Training
The multi-domain loader normalizes labels across datasets:
```yaml
# configs/splade_multidomain.yaml
run_name: splade_multidomain
base_model: distilbert/distilbert-... | [
0.042357951402664185,
-0.02839576080441475,
-0.04248945787549019,
-0.033764176070690155,
0.02938210964202881,
-0.0770428329706192,
0.09716922044754028,
0.07044865936040878,
0.035244181752204895,
-0.044844917953014374,
0.003260418539866805,
0.04512689635157585,
0.052171822637319565,
-0.0279... |
50be4875-733a-5b8e-b787-e1c4ed40238d | /documentation/embeddings/nomic/ | Nomic | Embed Images | documentation | true | Nomic > Embed Images
`embed.image` accepts file paths, URLs, or Pillow images. Image-to-text scores are much lower than text-to-text scores, so store image vectors in a collection of their own instead of mixing them with text documents. This example embeds images and searches them with a text query:
```python
client.... | [
0.042006365954875946,
-0.04418332129716873,
-0.029266778379678726,
-0.061497997492551804,
0.04847215116024017,
-0.06285180151462555,
0.07001079618930817,
0.04470894858241081,
0.031031733378767967,
-0.009010081179440022,
0.03325749561190605,
0.03138165920972824,
0.0467008613049984,
-0.01268... |
50c96e5e-62a2-57e8-8d1a-8fc710008152 | /documentation/tutorials-build-essentials/video-anomaly-edge-part-1/ | Video Anomaly Detection Part 1: Architecture, Twelve Labs, and NVIDIA VSS | Prerequisites | documentation | false | Video Anomaly Detection Part 1: Architecture, Twelve Labs, and NVIDIA VSS > Prerequisites
- **Twelve Labs API key** (required for cloud re-analysis, search, and Q&A): the [Free plan](https://www.twelvelabs.io/pricing) includes 600 minutes of video and needs no credit card. Larger baselines need a paid plan. See [Authe... | [
0.03433644026517868,
-0.023856867104768753,
-0.07727790623903275,
-0.025886479765176773,
0.06730227917432785,
-0.023247061297297478,
0.04486006498336792,
0.023994332179427147,
0.08917298913002014,
-0.027328891679644585,
0.059878233820199966,
-0.0062623945996165276,
0.05946708470582962,
-0.... |
50d972ba-43cf-5551-bc31-4f842da572de | /articles/vector-similarity-beyond-search/ | Vector Similarity: Going Beyond Full-Text Search | article | false | Vector Similarity: Going Beyond Full-Text Search
When making use of unstructured data, there are traditional go-to solutions that are well-known for developers:
- **Full-text search** when you need to find documents that contain a particular word or phrase.
- **[Vector search](/documentation/overview/vector-search/)*... | [
0.06975886970758438,
-0.06679196655750275,
-0.09453001618385315,
-0.04214581847190857,
0.05948923900723457,
-0.020209303125739098,
0.06805125623941422,
-0.01937374286353588,
0.04790075495839119,
-0.03436630219221115,
0.034698497503995895,
0.043455030769109726,
0.05836038663983345,
-0.01289... | |
50dc63eb-1731-5721-a9c0-ed2ade73e1ed | /documentation/ops-configuration/administration/ | Administration | Strict Mode > Maximum Batch Size When Searching | documentation | true | Administration > Strict Mode > Maximum Batch Size When Searching
Sending very large search batches can create internal congestion.
Setting `search_max_batchsize` caps the maximum number of searches in a single batch request.
```python
from qdrant_client import QdrantClient, models
client = QdrantClient(url="http://... | [
-0.015108130872249603,
-0.0370495468378067,
-0.06085945665836334,
-0.027082806453108788,
0.04786709323525429,
-0.014788158237934113,
0.04447639361023903,
0.04400179907679558,
0.061601970344781876,
-0.018251711502671242,
0.027809204533696175,
0.0369049534201622,
0.06189746409654617,
-0.0619... |
50ee80a3-bb4e-57fd-8e1c-69f8c14be3b8 | /documentation/private-cloud/qdrant-cluster-management/ | Managing a Cluster | Automatic Shard Rebalancing | documentation | true | Managing a Cluster > Automatic Shard Rebalancing
Qdrant Private Cloud supports automatic shard rebalancing. Shards are continuously redistributed across the available nodes to keep data evenly distributed, including when you scale up or down the number of nodes in a cluster. Automatic rebalancing applies to the whole ... | [
0.024466093629598618,
-0.023870104923844337,
-0.043412644416093826,
-0.055333659052848816,
0.06126433610916138,
-0.0603160634636879,
0.04864812269806862,
0.0382814034819603,
0.07198971509933472,
-0.03955001011490822,
0.04041391238570213,
0.019342748448252678,
0.035801053047180176,
-0.03112... |
50eecef1-758c-5f87-8e3a-42aeacc03d5f | /documentation/frameworks/mastra/ | Mastra | Methods > query() | documentation | false | Mastra > Methods > query()
| Name | Type | Description | Default Value |
|------------|-------------------------|---------------------------------------------|--------------|
| `vector` | `number[]` | Query vector to find similar vectors | ... | [
0.051197826862335205,
-0.06319551914930344,
-0.031061483547091484,
-0.003699274966493249,
0.04209969937801361,
-0.012889672070741653,
0.06457021832466125,
0.034638721495866776,
0.06307221204042435,
0.0012084938352927566,
0.025942351669073105,
0.05703257769346237,
0.03434572368860245,
-0.02... |
510ea9fe-4841-5e82-9518-78e182769276 | /documentation/tutorials-basics/search-beginners/ | Semantic Search 101 | documentation | false | Semantic Search 101
| Time: 5 - 15 min | Level: Beginner | | [](https://githubtocolab.com/qdrant/examples/blob/master/semantic-search-in-5-minutes/semantic_search_in_5_minutes.ipynb) |
| --- | ----------- | ----------- |----------- |
> There... | [
0.06060867756605148,
-0.06941770762205124,
-0.02991477958858013,
-0.025944411754608154,
0.07501152157783508,
-0.022716758772730827,
0.04203847795724869,
0.023401297628879547,
0.06216597557067871,
-0.03663548082113266,
0.04731884226202965,
0.01289046835154295,
0.06945966929197311,
-0.058253... | |
514f4d12-d8ee-5d76-ab67-cd5685387581 | /documentation/tutorials-search-engineering/collaborative-filtering/ | Collaborative Filtering | Implementation > Prepare the data | documentation | true | Collaborative Filtering > Implementation > Prepare the data
This tutorial uses the [MovieLens Small](https://files.grouplens.org/datasets/movielens/ml-latest-small.zip) dataset: about 100,000 ratings of 9,700 movies by 610 users. Download it and unpack it into a `data` folder:
```bash
wget https://files.grouplens.org... | [
0.0631498470902443,
-0.02935674600303173,
-0.020418640226125717,
-0.03292083367705345,
0.09369711577892303,
-0.048902031034231186,
0.018504196777939796,
0.10202813148498535,
0.04505021125078201,
-0.003933340776711702,
0.02382805198431015,
0.0032434165477752686,
0.0748801976442337,
-0.05343... |
517c42e9-1cd0-5b36-b6b7-6f5aa3146744 | /documentation/fastembed/fastembed-rerankers/ | Reranking with FastEmbed | Embed & index data for the first-stage retrieval | documentation | true | Reranking with FastEmbed > Embed & index data for the first-stage retrieval
We will vectorize a toy movie description dataset using the `all-MiniLM-L6-v2` model and save the embeddings in Qdrant for first-stage retrieval.
Then, we will use a cross-encoder reranking model to rerank a small subset of data retrieved in ... | [
0.07901439815759659,
-0.04224561154842377,
-0.012733143754303455,
-0.03309875726699829,
0.08180546760559082,
-0.04691319167613983,
0.06823676824569702,
0.053201138973236084,
0.060844868421554565,
0.008747940883040428,
0.02844630368053913,
0.012366660870611668,
0.06586050242185593,
-0.05261... |
51939531-c0b9-5eba-b2bb-ece43ed452f9 | /documentation/scaling/vertical-scaling/ | Vertical Scaling | Best Practices | documentation | false | Vertical Scaling > Best Practices
- Load test before scaling down RAM. Cache eviction after downsizing can cause a latency regression.
- Keep RAM usage below 80%. Memory pressure in Qdrant causes a performance cliff, not a gradual slowdown.
- Set the replication factor to two or higher before resizing in Qdrant Cloud.... | [
0.049704838544130325,
-0.037421900779008865,
-0.011792183853685856,
-0.05741393193602562,
0.0815480500459671,
-0.03358624875545502,
0.05290025472640991,
0.010548730380833149,
0.10600410401821136,
-0.06682752817869186,
-0.003447322640568018,
0.00012272200547158718,
0.02381797693669796,
-0.0... |
519539d4-2f89-5355-9173-af352c07ae45 | /articles/late-interaction-models/ | Late Interaction Retrieval with Dense Token Embeddings | Comparative Analysis of Dense vs. Late Interaction Models | article | false | Late Interaction Retrieval with Dense Token Embeddings > Comparative Analysis of Dense vs. Late Interaction Models
The retrieval quality speaks for itself, but there are other important factors to consider.
The traditional dense embedding models we tested are less complex than late interaction or sparse models. With ... | [
0.0536593459546566,
-0.012005588039755821,
-0.044849734753370285,
-0.062492575496435165,
0.07172577828168869,
-0.05219510570168495,
0.056330323219299316,
0.08128752559423447,
0.09699384868144989,
-0.02819296345114708,
0.03538297116756439,
0.007870839908719063,
0.02315550297498703,
-0.01471... |
51ae393e-a133-5a27-ba7a-bf11c7fcd662 | /documentation/search/filtering/ | Filtering | Filtering Conditions > Has ID | documentation | true | Filtering > Filtering Conditions > Has ID
This type of query is not related to payload, but can be very useful in some situations.
For example, the user could mark some specific search results as irrelevant, or we want to search only among the specified points.
```python
client.scroll(
collection_name="{collectio... | [
0.011997129768133163,
-0.05099400877952576,
-0.06611587107181549,
-0.02504643052816391,
0.0569741427898407,
-0.013196681626141071,
0.036552634090185165,
0.03557788208127022,
0.09736589342355728,
-0.01814955286681652,
0.02452436089515686,
0.03084084577858448,
0.03854883462190628,
-0.0002568... |
51cff3bf-18fa-5af2-ac6a-611f81f2db24 | /documentation/faq/qdrant-fundamentals/ | Qdrant Fundamentals | Collections > Should I use named vectors or separate collections for different embedding models? | documentation | false | Qdrant Fundamentals > Collections > Should I use named vectors or separate collections for different embedding models?
Use **named vectors** when the data you're embedding shares the same payload structure and you want to query across vector spaces in a single request (for example, combining a dense text vector with a... | [
0.051431987434625626,
-0.023678194731473923,
-0.030401039868593216,
-0.028451353311538696,
0.04706325754523277,
-0.060099080204963684,
0.06508856266736984,
0.0338604561984539,
0.05345886945724487,
-0.02833796665072441,
0.02991177886724472,
0.004909869283437729,
0.02959608845412731,
-0.0102... |
51ec1920-d0b0-5504-b82e-98cfd87c40c8 | /course/essentials/day-7/haystack/ | Integrating with Haystack | Haystack Movie Recommendation Assistant > Advanced Query Handling | course | false | Integrating with Haystack > Haystack Movie Recommendation Assistant > Advanced Query Handling
The system excels at processing sophisticated queries by:
- **Natural Language Understanding**: Interpreting queries like "highly-rated action movie about car racing"
- **Multi-Criteria Filtering**: Combining genre, rating, ... | [
0.022722871974110603,
-0.046401526778936386,
-0.07992000877857208,
0.006695740856230259,
0.07625272125005722,
-0.03744017705321312,
0.012453502975404263,
0.0028438486624509096,
0.06325419992208481,
-0.010199464857578278,
0.031102702021598816,
0.00781518779695034,
0.07777302712202072,
-0.04... |
520d8243-9697-5180-9b0e-629c44b456f3 | /documentation/embeddings/ollama/ | Ollama | Embed Queries | documentation | true | Ollama > Embed Queries
Ollama sends text to the model as is. Qwen3 Embedding expects an instruction in front of each query, and none in front of documents:
```python
TASK = "Given a web search query, retrieve relevant passages that answer the query"
query_vector = ollama_client.embed(
model=MODEL,
input=f"In... | [
0.039009395986795425,
-0.06504099071025848,
-0.04551292583346367,
-0.02198207937180996,
0.08878224343061447,
-0.030927389860153198,
0.0690498873591423,
0.049944691359996796,
0.019885599613189697,
-0.006237410474568605,
0.02713088132441044,
0.01814734935760498,
0.011149036698043346,
-0.0296... |
520f711b-9ae2-5696-a29d-e6fe3521ddd0 | /documentation/cloud/cluster-upgrades/ | Update Clusters | documentation | false | Update Clusters
As soon as a new Qdrant version is available. Qdrant Cloud will show you an update notification in the Cluster list and on the Cluster details page.
To update to a new version, go to the Cluster Details page, choose the new version from the version dropdown and click **Update**.
If you are several ve... | [
-0.002407070714980364,
-0.04193073883652687,
-0.03825761750340462,
-0.023471975699067116,
0.07309047132730484,
-0.06070515885949135,
0.04798189550638199,
0.02600422315299511,
0.06507689505815506,
-0.05729090794920921,
0.02343634143471718,
0.03131171315908432,
0.048462145030498505,
-0.05239... | |
521525fa-c9dd-51f3-a5d9-ff154ae82001 | /documentation/frameworks/spring-ai/ | Spring AI | documentation | true | Spring AI
[Spring AI](https://docs.spring.io/spring-ai/reference/) is a Java framework that provides a [Spring-friendly](https://spring.io/) API and abstractions for developing AI applications.
Qdrant is available as supported vector database for use within your Spring AI projects.
## Installation
You can find the ... | [
0.037721216678619385,
-0.01766319014132023,
-0.03320638835430145,
-0.06788479536771774,
0.03707481548190117,
-0.02031506970524788,
0.08345083892345428,
0.07419701665639877,
0.06259990483522415,
-0.054109349846839905,
0.06343904882669449,
0.013876073062419891,
0.024362288415431976,
-0.02811... | |
52265000-ad81-5899-86a7-d1fd5b896c4a | /articles/agentic-rag/ | What is Agentic RAG? Building Agents with Qdrant | Which Framework is Best? > LangGraph | article | true | What is Agentic RAG? Building Agents with Qdrant > Which Framework is Best? > LangGraph
- **Persistence** - the state of the workflow graph is stored as a checkpoint. That happens at each so-called super-step
(which is a single sequential node of a graph). It enables replying certain steps of the workflow, fault-tol... | [
0.020277023315429688,
-0.029426369816064835,
-0.0608823262155056,
-0.026468688622117043,
0.10807487368583679,
-0.06150912120938301,
0.06312151998281479,
0.006658516824245453,
0.05369696393609047,
-0.03648771345615387,
0.030132291838526726,
0.004789059981703758,
0.07667887210845947,
-0.0262... |
522a6eba-73b3-5be7-8d37-17068854b6ec | /documentation/scaling/vertical-scaling/ | Vertical Scaling | When to Scale Vertically | documentation | false | Vertical Scaling > When to Scale Vertically
Scale vertically when your current node resources are insufficient but your workload doesn't yet require distribution:
- RAM usage is approaching 80% of available memory. Beyond this threshold, the operating system starts evicting pages from cache, which causes a sharp perf... | [
0.026947105303406715,
-0.03576739877462387,
-0.02300686202943325,
-0.026062874123454094,
0.06888990104198456,
-0.01589803956449032,
0.03927931934595108,
0.022986426949501038,
0.11807084828615189,
-0.041976213455200195,
0.005127209238708019,
-0.013158949092030525,
0.017347106710076332,
-0.0... |
523d74b7-6d0b-5e92-b93b-1987340cdb42 | /documentation/migration-guidance/diagnosing-discrepancies/ | Diagnosing Discrepancies | Vendor-Specific Gotchas | documentation | false | Diagnosing Discrepancies > Vendor-Specific Gotchas
* **Schema strictness:** Milvus enforces schema on write; Qdrant is schema-flexible. Verify that schema-less flexibility didn't cause payload fields to drift during migration.
* **Partition mapping:** Milvus partitions can map to Qdrant collections or payload filters.... | [
0.05898057669401169,
-0.044395092874765396,
-0.059199556708335876,
-0.016597867012023926,
0.020993242040276527,
-0.01491284929215908,
0.07535074651241302,
0.030470624566078186,
0.06411321461200714,
-0.018220651894807816,
0.020671993494033813,
-0.0034298913087695837,
0.04676640406250954,
-0... |
5243fa54-cab9-55e7-8294-b7e42ba65319 | /documentation/tutorials-search-engineering/ann-recall/ | Measuring ANN Recall | documentation | false | Measuring ANN Recall
| Time: 15 min | Level: Beginner | | |
|--------------|---------------------|--|----|
This tutorial focuses on **ANN recall**: how closely approximate nearest-neighbor (ANN) search matches exact kNN search, measured with `recall@k`.
**Prerequisites.** A Qdrant collection populated with your ... | [
0.016374075785279274,
-0.05831323191523552,
-0.027819357812404633,
-0.05802550166845322,
0.07859063893556595,
-0.019079726189374924,
0.019623523578047752,
0.062088556587696075,
0.10381356626749039,
-0.032207097858190536,
0.04664119333028793,
0.01989835686981678,
0.05256109684705734,
-0.022... | |
5258632f-acf1-5ba6-8625-4a25ff30e408 | /documentation/manage-data/payload/ | Payload | Facet counts | documentation | true | Payload > Facet counts
*Available as of v1.12.0*
Faceting is a special counting technique that can be used for various purposes:
- Know which unique values exist for a payload key.
- Know the number of points that contain each unique value.
- Know how restrictive a filter would become by matching a specific value.
S... | [
0.02581160143017769,
-0.06106451898813248,
-0.03783241659402847,
-0.053017426282167435,
0.05863286182284355,
-0.02192123420536518,
0.02673555724322796,
0.02448820322751999,
0.09728173911571503,
-0.03741017356514931,
0.04540316015481949,
-0.005294193979352713,
0.0491141676902771,
-0.0281078... |
5277c702-db22-585e-8d19-3e8736399685 | /documentation/data-synchronization/with-postgres/ | Keeping Postgres and Qdrant in Sync | Tier 3: Change Data Capture with Debezium + Redpanda > When to Use This | documentation | false | Keeping Postgres and Qdrant in Sync > Tier 3: Change Data Capture with Debezium + Redpanda > When to Use This
- High write throughput systems where the outbox table would become a bottleneck
- When multiple downstream consumers need to react to changes (not just Qdrant)
- When application code must be fully decoupled ... | [
0.005026471801102161,
-0.018769266083836555,
-0.046805646270513535,
-0.041759390383958817,
0.032592155039310455,
-0.05294490605592728,
0.06836686283349991,
0.027369696646928787,
0.07724155485630035,
-0.019785841926932335,
0.03173103928565979,
0.023556161671876907,
0.04022456705570221,
-0.0... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.