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 |
|---|---|---|---|---|---|---|---|
cc1c8c53-37d7-5b22-8782-0086cd2b1a54 | /articles/what-is-a-vector-database/ | What is a Vector Database? | Multitenancy: Data Isolation for Multi-Tenant Architectures | article | true | What is a Vector Database? > Multitenancy: Data Isolation for Multi-Tenant Architectures
Sharding efficiently distributes data across nodes, while replication guarantees redundancy and fault tolerance. But what happens when you’ve got multiple clients or user groups, and you need to keep their data isolated within the... | [
0.01694578118622303,
-0.010914408601820469,
-0.08573465049266815,
-0.011610087007284164,
0.05147400498390198,
-0.021525852382183075,
0.0838642492890358,
0.033938050270080566,
0.05350213125348091,
-0.02099403366446495,
0.03115658648312092,
0.0330745093524456,
0.08096175640821457,
-0.0060679... |
cc368ead-1f43-5c18-ae88-74bdd9183150 | /documentation/tutorials-basics/search-beginners-local/ | Semantic Search 101 | 5. Upload Data to Collection | documentation | true | Semantic Search 101 > 5. Upload Data to Collection
Tell the database to upload `documents` to the `my_books` collection. This will give each record an id and a payload. The payload is just the metadata from the dataset.
```python
client.upload_points(
collection_name="my_books",
points=[
models.PointS... | [
-0.010865381918847561,
-0.023074522614479065,
-0.049159981310367584,
-0.03823798894882202,
0.08803239464759827,
-0.013780533336102962,
0.03275981545448303,
0.04226934164762497,
0.05496629700064659,
-0.023447111248970032,
0.002114176284521818,
0.047311317175626755,
0.06505067646503448,
-0.0... |
cc659134-1ce5-5483-8d12-5fb5e12a7d45 | /documentation/overview/what-is-qdrant/ | What is Qdrant? | Next Steps | documentation | false | What is Qdrant? > Next Steps
Now that you know more about vector databases and Qdrant, you are ready to get started with one
of our tutorials. If you've never used a vector database, go ahead and jump straight into
the **Getting Started** section. Conversely, if you are a seasoned developer in these
technology, jum... | [
0.032138653099536896,
-0.07023083418607712,
-0.04217541590332985,
-0.033594731241464615,
0.02232367731630802,
-0.05261072888970375,
0.07743305712938309,
0.02127157896757126,
0.07217004895210266,
-0.029664786532521248,
0.025245103985071182,
0.028255056589841843,
0.049677636474370956,
-0.031... |
cc762601-a643-5efb-bb6d-2541feb653f0 | /documentation/examples/rag-chatbot-red-hat-openshift-haystack/ | Private Chatbot for Interactive Learning | Implementation | documentation | false | Private Chatbot for Interactive Learning > Implementation
We will first create an indexing pipeline to add documents to the system.
Then, the search pipeline will retrieve relevant data from our documents.
After the pipelines are tested, we will export them to YAML files. | [
-0.0055326600559055805,
-0.01449523400515318,
-0.046799931675195694,
-0.028484825044870377,
0.06408035010099411,
-0.021059930324554443,
0.039030350744724274,
0.029017051681876183,
0.03674862161278725,
-0.0099119171500206,
0.017858019098639488,
0.060468416661024094,
0.06741934269666672,
-0.... |
cc7bd824-e69b-5734-9e9e-00e70c0f864f | /documentation/ops-configuration/configuration/ | Configuration | Configuration Files | documentation | true | Configuration > Configuration Files
To customize Qdrant, you can mount your configuration file in any of the following locations. This guide uses `.yaml` files, but Qdrant also supports other formats such as `.toml`, `.json`, and `.ini`.
1. **Main Configuration: `qdrant/config/config.yaml`**
Mount your custom `confi... | [
0.04740431532263756,
-0.0652516633272171,
-0.056523725390434265,
-0.04399610310792923,
0.0374479666352272,
-0.045950695872306824,
0.04061035066843033,
0.05045156553387642,
0.06197525933384895,
0.003786622080951929,
0.007835961878299713,
-0.008687012828886509,
0.04577379301190376,
-0.055721... |
ccaee316-2a15-5ac6-86b2-ee386a560029 | /articles/what-is-rag-in-ai/ | Understanding Retrieval-Augmented Generation (RAG) | RAG architecture | article | false | Understanding Retrieval-Augmented Generation (RAG) > RAG architecture
At its core, a RAG architecture includes the **retriever** and the **generator**. Let's start by understanding what each of these components does. | [
-0.03848249837756157,
-0.027038346976041794,
-0.0529344379901886,
-0.05808961018919945,
0.07778222858905792,
-0.03892242908477783,
0.025746500119566917,
0.06760932505130768,
0.02691335789859295,
-0.039643898606300354,
0.030564282089471817,
0.04496811330318451,
0.08126186579465866,
-0.00370... |
ccbf5e14-0622-55f5-880a-cf471488542a | /documentation/tutorials-search-engineering/branch-aware-search/ | Branch-Aware Search | Step 5: Change an Inherited File on the Branch | documentation | true | Branch-Aware Search > Step 5: Change an Inherited File on the Branch
`A` edits `pricing.md`, a file it inherited from `root`.
```text
root ●──●──● seq 0–2
└── A ●──● A seq 1: update pricing.md
```
Nothing new is needed. `update` looks up the version `A` currently sees, which is `root`'s, an... | [
0.04433858022093773,
-0.08585119247436523,
-0.04071739688515663,
-0.024239476770162582,
0.04527360200881958,
-0.029403286054730415,
-0.01251385360956192,
0.07547760754823685,
0.03917594999074936,
-0.0281792264431715,
0.029598690569400787,
0.05239419639110565,
0.07325054705142975,
-0.033482... |
ccc17a06-1d07-5aca-901f-435412f42b9c | /documentation/examples/graphrag-qdrant-neo4j/ | GraphRAG with Qdrant and Neo4j | Implementation > Ingestion > Ingesting Data to Neo4j | documentation | true | GraphRAG with Qdrant and Neo4j > Implementation > Ingestion > Ingesting Data to Neo4j
Each paragraph becomes a `Chunk` node, linked to the entities it mentions by `MENTIONS` relationships. The chunk's ID is what connects Neo4j to Qdrant.
```python
def ingest_to_neo4j(nodes, relationships, chunk_id):
"""
Inges... | [
0.046348150819540024,
-0.024816283956170082,
-0.07852395623922348,
-0.03933367133140564,
0.08213584870100021,
-0.05878026410937309,
0.033631954342126846,
0.09836076200008392,
0.053147442638874054,
-0.04440274089574814,
0.036410100758075714,
0.027695713564753532,
0.05723898857831955,
-0.015... |
ccd090b6-82dc-56c6-8376-4b14564539b0 | /articles/bm42/ | BM42: Attention-Based Sparse Embeddings for Hybrid Search | Why SPLADE is not always the answer | article | false | BM42: Attention-Based Sparse Embeddings for Hybrid Search > Why SPLADE is not always the answer
* **Expensive Token Expansion**: In order to compensate the tokenization issues, SPLADE uses *token expansion* technique. This means that we generate a set of similar tokens for each token in the query. There are a few prob... | [
0.05003125220537186,
-0.03142580762505531,
-0.04748997464776039,
-0.05666879191994667,
0.04798591509461403,
-0.0443606823682785,
0.062088411301374435,
0.03099605068564415,
0.07145846635103226,
-0.029497722163796425,
0.04285217449069023,
-0.004644227214157581,
0.04720716550946236,
-0.007008... |
cce6d22f-70e9-5888-b389-2e42ae4b2424 | /course/essentials/day-3/sparse-retrieval-demo/ | Demo: Keyword Search with Sparse Vectors | What You'll Learn | course | false | Demo: Keyword Search with Sparse Vectors > What You'll Learn
Day 3
Use sparse vectors for keywords-based text retrieval.
- Connection between Sparse Vectors & keywords-based retrieval
- Using BM25 in Qdrant
- Sparse Neural Retrieval
- Using SPLADE++ in Qdrant | [
0.05737373232841492,
-0.053629107773303986,
-0.03499549627304077,
-0.04727096110582352,
0.060749541968107224,
-0.06351236253976822,
0.060614146292209625,
0.004645568318665028,
0.048708997666835785,
-0.014646747149527073,
0.02202383428812027,
0.030170323327183723,
0.03788931667804718,
-0.00... |
ccf08a84-213c-5af4-a98a-800c97f1905d | /documentation/manage-data/indexing/ | Indexing | Payload Index > Create a Payload Index | documentation | true | Indexing > Payload Index > Create a Payload Index
To create a payload index for a field:
```python
client.create_payload_index(
collection_name="{collection_name}",
field_name="name_of_the_field_to_index",
field_schema=models.PayloadSchemaType.KEYWORD,
)
```
You can use dot notation to specify a nested f... | [
0.03146838769316673,
-0.010992554016411304,
-0.03669661656022072,
-0.06046605110168457,
0.04883350059390068,
-0.059182990342378616,
0.03592653200030327,
0.026368718594312668,
0.08875885605812073,
-0.03910159692168236,
0.039397936314344406,
0.03453126549720764,
0.03513544425368309,
-0.02523... |
ccf4cbc7-0aea-5b64-9b9c-ffc4d99e49ce | /course/beginners/module-4/decide-before-you-ingest/ | Decide Before You Ingest | Build in This Order | course | true | Decide Before You Ingest > Build in This Order
Create the collection, create its payload indexes, then ingest the points. When Qdrant builds the HNSW graph, it adds extra edges for fields that already have payload indexes, making filtered search faster. An index added later still filters correctly, but Qdrant must [re... | [
0.04131433740258217,
-0.035825278609991074,
-0.03864679113030434,
-0.018027277663350105,
0.085106261074543,
-0.04269583523273468,
0.07901252061128616,
0.07453334331512451,
0.020687473937869072,
-0.02635042555630207,
0.03139809891581535,
0.04627484455704689,
0.020472439005970955,
-0.0421169... |
ccfd9669-bf9d-51ad-bff3-4e9d6c6b3756 | /documentation/tutorials-operations/prevent-unoptimized-usage/ | Prevent Unoptimized Usage | Tradeoffs | documentation | false | Prevent Unoptimized Usage > Tradeoffs
The faster optimization and steady query latency might suggest `prevent_unoptimized` is always the right call. It is not the full story: by definition, `prevent_unoptimized` withholds points in segments that have not finished optimizing, from search results.
This means searches r... | [
0.025298459455370903,
-0.05696670338511467,
-0.05893555283546448,
-0.034971244633197784,
0.07336510717868805,
-0.032393962144851685,
0.0455593504011631,
0.029580846428871155,
0.07881606370210648,
-0.03153438866138458,
0.046268220990896225,
0.01432530116289854,
0.08145879954099655,
-0.01025... |
cd03d34a-af3c-55ce-ac09-5a6fc4267f7a | /documentation/fastembed/fastembed-rerankers/ | Reranking with FastEmbed | First-stage retrieval | documentation | true | Reranking with FastEmbed > First-stage retrieval
Let's see how relevant the results will be using only an `all-MiniLM-L6-v2`-based dense retriever.
```python
query = "A story about a strong historically significant female figure."
query_embedded = list(dense_embedding_model.query_embed(query))[0]
initial_retrieval =... | [
0.062142372131347656,
-0.04286772757768631,
-0.040143415331840515,
-0.022482113912701607,
0.08496492356061935,
-0.056276869028806686,
0.0533299595117569,
0.03171597421169281,
0.04796125367283821,
-0.005004050210118294,
0.0380488745868206,
0.012180459685623646,
0.06829946488142014,
-0.02365... |
cd0462e1-a136-5754-a24b-55810c6c7c08 | /articles/filtered-vector-search-acorn/ | Filtered Vector Search: What ACORN Fixes, and What Fixes ACORN | The Graph Qdrant Builds Instead | article | false | Filtered Vector Search: What ACORN Fixes, and What Fixes ACORN > The Graph Qdrant Builds Instead
[Filterable HNSW](/articles/filterable-hnsw/), which our co-founder Andrey Vasnetsov described in 2019, builds the repair into the index. When a payload field, the metadata attached to each point, is [indexed](/documentati... | [
0.02267622947692871,
-0.044543541967868805,
-0.048298005014657974,
-0.002858072752133012,
0.0740017518401146,
-0.023142384365200996,
0.05244047939777374,
0.04914120212197304,
0.05824267864227295,
-0.016865992918610573,
0.06772513687610626,
0.02507932297885418,
0.03773224353790283,
-0.01135... |
cd0dd33d-637f-58f4-b513-52e3cbd85e57 | /documentation/fastembed/fastembed-postprocessing/ | Multi-Vector Postprocessing | Using MUVERA in FastEmbed | documentation | false | Multi-Vector Postprocessing > Using MUVERA in FastEmbed
This tutorial demonstrates using MUVERA for fast retrieval with ColBERT reranking on a toy dataset. If you're new to
multi-vector embeddings, we recommend first reading [How to Generate ColBERT Multivectors with
FastEmbed](/documentation/fastembed/fastembed-col... | [
0.05370039492845535,
-0.020975913852453232,
-0.0350661426782608,
-0.07799208164215088,
0.08177356421947479,
-0.029474643990397453,
0.07002685964107513,
0.01878378540277481,
0.042688317596912384,
0.023673085495829582,
0.012903803028166294,
0.043820783495903015,
0.03315499424934387,
-0.03395... |
cd1899af-54d5-5030-8928-fdc6c52b7974 | /course/essentials/day-4/pitstop-project/ | Project: Quantization Performance Optimization | Your Mission | course | false | Project: Quantization Performance Optimization > Your Mission
Transform your search engine from previous days into a production-ready system by implementing quantization optimization. You'll test different quantization methods, measure performance impacts, and tune the oversampling + rescoring pipeline for optimal res... | [
0.005114120431244373,
-0.04236948490142822,
-0.022240985184907913,
-0.009385105222463608,
0.07798600196838379,
-0.025312330573797226,
0.027016770094633102,
0.01004071906208992,
0.09976626932621002,
-0.014613915234804153,
0.026710664853453636,
0.012380162253975868,
0.05756327509880066,
-0.0... |
cd296290-d55c-5685-a0fb-d53b2ff69fe1 | /documentation/tutorials-operations/create-snapshot/ | Snapshots | Prerequisites > Upload the dataset | documentation | true | Snapshots > Prerequisites > Upload the dataset
Calculating the embeddings is usually a bottleneck of the vector search pipelines, but we are happy to have them in place already. Since the goal of this tutorial is to show how to create a snapshot, **we are going to upload only a small part of the dataset**.
```python
... | [
0.0029573170468211174,
-0.03842362016439438,
-0.04607043415307999,
-0.044425781816244125,
0.0520927719771862,
-0.02698720619082451,
0.05948534235358238,
0.027791758999228477,
0.0784730613231659,
-0.03643651306629181,
0.029088422656059265,
0.03588588908314705,
0.06042404845356941,
-0.041955... |
cd3953f8-d91d-546d-91b7-7afc86b76d55 | /documentation/faq/qdrant-fundamentals/ | Qdrant Fundamentals | Vectors > Can Qdrant generate vector embeddings? | documentation | false | Qdrant Fundamentals > Vectors > Can Qdrant generate vector embeddings?
Yes, if you're using Qdrant Cloud, you can generate embeddings with [**Qdrant Cloud Inference**](/documentation/inference/). It lets you embed, store, and index your data in a single API call, so you don't need a separate inference service or embed... | [
0.0488395169377327,
-0.017030443996191025,
-0.013660192489624023,
-0.06216258183121681,
0.07038912177085876,
-0.037006676197052,
0.09789116680622101,
0.027808140963315964,
0.05786500498652458,
-0.031444747000932693,
0.029272805899381638,
0.02088683284819126,
0.043397996574640274,
-0.042545... |
cd49c5a6-c222-5899-873d-3c5591bb5e3c | /documentation/search/search/ | Search | Search API > Payload and Vector in the Result | documentation | true | Search > Search API > Payload and Vector in the Result
By default, retrieval methods do not return any stored information such as
payload and vectors. Additional parameters `with_vectors` and `with_payload`
alter this behavior.
Example:
```python
client.query_points(
collection_name="{collection_name}",
quer... | [
0.02635163813829422,
-0.032658278942108154,
-0.01860600709915161,
-0.04162753000855446,
0.060500700026750565,
-0.03353211283683777,
0.032914865761995316,
0.028367262333631516,
0.058426596224308014,
-0.028963034972548485,
0.04936440289020538,
0.04611124098300934,
0.027458498254418373,
0.010... |
cd663ce2-7164-51ac-bdf4-dac1b07fe9bc | /documentation/data-management/cocoindex/ | CocoIndex | documentation | true | CocoIndex
[CocoIndex](https://cocoindex.io) is a high performance ETL framework to transform data for AI, with real-time incremental processing.
Qdrant is available as a native built-in vector database to store and retrieve embeddings.
Install CocoIndex:
```bash
pip install -U cocoindex
```
Install Postgres with [... | [
0.03416692465543747,
-0.05230691283941269,
-0.030702395364642143,
-0.047414712607860565,
0.06246434897184372,
-0.04192216694355011,
0.054718319326639175,
0.021062877029180527,
0.05203545093536377,
-0.07143445312976837,
0.0162193663418293,
0.038296353071928024,
0.03664849326014519,
-0.05486... | |
cd6a9052-327a-502c-952d-b0614abbb164 | /documentation/private-cloud/changelog/ | Changelog | 1.7.3 (2025-08-04) | documentation | false | Changelog > 1.7.3 (2025-08-04)
| Component | Version |
|------------------------|---------|
| qdrant-kubernetes-api | v1.17.2 |
| operator | 2.6.8 |
| qdrant-cluster-manager | v0.3.7 |
* Configurable max_request_size_mb | [
0.03398337587714195,
-0.009484702721238136,
-0.04820411652326584,
-0.044131603091955185,
0.06818164885044098,
-0.06052837520837784,
0.05593708157539368,
0.056202054023742676,
0.04845713451504707,
-0.036084286868572235,
0.019835811108350754,
0.03439919650554657,
0.06362545490264893,
-0.0692... |
cd778c66-1a88-547f-80a6-b461f842f458 | /articles/distance-based-exploration/ | Data Exploration with Qdrant's Distance Matrix API | Conclusion | article | false | Data Exploration with Qdrant's Distance Matrix API > Conclusion
Vector similarity goes beyond looking up the nearest neighbors. It also gives you a tool for data exploration.
Many algorithms can construct human-readable data representations, and Qdrant supplies the neighbor data they need.
The Qdrant Web UI includes ... | [
0.0636085793375969,
-0.0719580128788948,
-0.047656577080488205,
-0.018078939989209175,
0.0697961151599884,
-0.012499786913394928,
0.12168639898300171,
0.03465145081281662,
0.04387744888663292,
-0.03908514976501465,
0.06212405487895012,
0.029167858883738518,
0.06262696534395218,
0.022872535... |
cd794d79-55c2-5b1d-8f9f-5a017364f511 | /course/essentials/day-0/pitstop-project/ | Project: Building Your First Vector Search System | Your Mission | course | false | Project: Building Your First Vector Search System > Your Mission
Day 0
Time to apply what you've learned. You'll create a complete, working vector search system from scratch.
Build a functional vector search system that demonstrates the core concepts: collections, points, similarity search, and filtering. You'll des... | [
0.03314243257045746,
-0.053428228944540024,
-0.018765408545732498,
0.016554171219468117,
0.08372783660888672,
-0.02378849685192108,
0.022661982104182243,
0.047977108508348465,
0.09311468154191971,
-0.026645438745617867,
0.03335612267255783,
0.030285166576504707,
0.0709657296538353,
-0.0397... |
cd92fabf-84aa-54e3-9b9b-543ac26d2a56 | /documentation/production-checklist/ | Production Checklist | 2. Security | documentation | false | Production Checklist > 2. Security
Lock down access before exposing your instance to any network. By default, self-hosted open source Qdrant instances are open to all interfaces without any authentication. On Qdrant Cloud, security features are enabled by default.
- **[Set Up an API Key](/documentation/security/#auth... | [
0.015282869338989258,
-0.05603349581360817,
-0.05674159526824951,
-0.035936255007982254,
0.043897129595279694,
-0.03498923406004906,
0.042070090770721436,
0.03762011229991913,
0.08365181088447571,
-0.02273888699710369,
0.013234819285571575,
-0.0075165098533034325,
0.05705151706933975,
-0.0... |
cda03627-f7b3-5379-b85d-2fd1ba53dc0e | /course/multi-vector-search/module-2/how-colpali-works/ | How ColPali Models Work | course | false | How ColPali Models Work
Module 2
ColPali extends the late interaction paradigm from text to visual documents. It can process PDFs, images, and scanned documents, generating multi-vector representations that capture both textual and visual information.
Understanding ColPali's architecture helps you leverage its full ... | [
0.006740550510585308,
-0.022216733545064926,
-0.04279917851090431,
-0.06286519765853882,
0.08194376528263092,
-0.010692541487514973,
0.08940571546554565,
0.06787469983100891,
0.022457517683506012,
-0.014740580692887306,
0.014330240897834301,
0.010729320347309113,
0.04260222613811493,
-0.01... | |
cdb511b6-5434-52c3-b344-2112e523cbb4 | /documentation/deploy-intro/ | Deploy Qdrant | Hybrid Cloud | documentation | false | Deploy Qdrant > Hybrid Cloud
- [Hybrid Cloud](/documentation/hybrid-cloud/index.md) — Deploy Qdrant in your own Kubernetes cluster while managing it through Qdrant Cloud.
- [Setup Hybrid Cloud](/documentation/hybrid-cloud/hybrid-cloud-setup/index.md) — Install and connect the Qdrant Kubernetes Operator to Qdrant Cloud... | [
0.008861806243658066,
-0.02664819546043873,
-0.060254428535699844,
-0.059385593980550766,
0.0751025527715683,
-0.024466006085276604,
0.05326969549059868,
0.031970683485269547,
0.02606252022087574,
-0.021853560581803322,
0.035702940076589584,
-0.002905517118051648,
0.049011170864105225,
-0.... |
cdc50575-71bf-5a46-9e5e-389a84cc92dd | /course/beginners/module-5/module-overview/ | Module Overview | course | false | Module Overview
Module 5
Apply every concept from Modules 1 through 4 in a single end-to-end system: ingest daily news, transcripts, and satellite imagery about your suppliers, cluster them into risk themes, and query all of it from one collection.
**Follow-along code**: [Module 5 notebook](https://github.com/qdrant... | [
0.05753137543797493,
-0.01477549783885479,
-0.053171586245298386,
-0.03089081309735775,
0.07736776769161224,
-0.04527617245912552,
0.03110187128186226,
0.056365787982940674,
0.0739150270819664,
-0.014330873265862465,
0.009154981933534145,
0.020428191870450974,
0.05096319690346718,
-0.03019... | |
cdc8b8cd-0c18-5bf0-9050-0c3b093465ff | /course/beginners/module-2/module-overview/ | Module Overview | course | false | Module Overview
Module 2
Understand collections, points, vectors, payloads, and the HNSW index, and move from theory to actual system design in Qdrant.
**Follow-along code**: [Module 2 notebook](https://github.com/qdrant/examples/blob/master/course/beginners/Module2.ipynb)
#### Overview
> Module 1 explained why se... | [
0.05058179050683975,
-0.042384229600429535,
-0.04077863320708275,
-0.0016059235204011202,
0.06209548935294151,
-0.03155980631709099,
0.042460232973098755,
0.04389413818717003,
0.09083876013755798,
-0.010586697608232498,
0.02377302572131157,
0.03652338683605194,
0.046977780759334564,
-0.019... | |
cdcf4a8c-99b7-5726-86c3-b88337cc7e03 | /documentation/edge/edge-vs-qdrant-cluster/ | Edge vs. Qdrant Cluster | Operations | documentation | false | Edge vs. Qdrant Cluster > Operations
How data gets indexed, optimized, and kept available.
| | Qdrant Edge | Qdrant Server |
| --- | --- | --- |
| **Optimization** | Manual, by calling `optimize()`; runs synchronously | Continuous background optimizer |
| **HNSW indexing** | Manual indexing for large shards via `opti... | [
0.0006286980933509767,
-0.015428527258336544,
-0.06673844903707504,
-0.03006420098245144,
0.08924636989831924,
-0.04835820570588112,
0.07710954546928406,
0.009868962690234184,
0.06551818549633026,
-0.03318607434630394,
0.046643756330013275,
0.02062218263745308,
0.03411471098661423,
-0.0212... |
cdd2d711-fc3e-536b-a307-d429e9966d90 | /documentation/tutorials-build-essentials/video-anomaly-edge-part-3/ | Video Anomaly Detection Part 3: Scoring, Governance, and Deployment | documentation | false | Video Anomaly Detection Part 3: Scoring, Governance, and Deployment
| Time: 90 min | Level: Advanced | Output: [GitHub](https://github.com/qdrant/video-anomaly-edge) |
| --- | ----------- | ----------- |
*This is Part 3 of a 3-part series on building real-time video anomaly detection from edge to cloud.*
**Series:**... | [
0.03822368010878563,
-0.027010925114154816,
-0.0892658606171608,
-0.04066399857401848,
0.07016643136739731,
-0.011532957665622234,
0.040644217282533646,
0.02561277151107788,
0.07055451720952988,
-0.028746603056788445,
0.07056110352277756,
0.012568317353725433,
0.03863529860973358,
-0.02883... | |
cde13aa6-96ec-570a-ba48-41777547b36a | /documentation/migrate-to-qdrant/from-milvus/ | From Milvus | What You Need from Milvus | documentation | false | From Milvus > What You Need from Milvus
- **Milvus URL** — the gRPC endpoint of your Milvus instance
- **Collection name** — the collection to migrate
- **API key** — if using Zilliz Cloud or authenticated Milvus | [
0.02937239781022072,
-0.043231040239334106,
-0.06392410397529602,
-0.047634292393922806,
0.06599141657352448,
-0.007751223165541887,
0.04785866290330887,
0.04833243787288666,
0.036987096071243286,
-0.03335385397076607,
0.0312553271651268,
0.0009491490782238543,
0.029009558260440826,
-0.051... |
cde19258-2a6c-5b26-9971-dd363e7e51aa | /documentation/embeddings/upstage/ | Upstage | Embed Queries | documentation | true | Upstage > Embed Queries
Embed queries with the query model. The query model is tuned for short questions and search terms, and the passage model for the longer texts that answer them. Both write to the same vector space, so a query vector can be compared directly with the passage vectors in the collection:
```python
... | [
0.017277933657169342,
-0.05004359409213066,
-0.038949351757764816,
-0.037317950278520584,
0.050945475697517395,
-0.012738259509205818,
0.07461821287870407,
0.03891916945576668,
0.05540010333061218,
0.0016969647258520126,
0.023712443187832832,
0.015731316059827805,
0.04158055782318115,
-0.0... |
cde77156-0dd3-59dd-a336-05b2d629e5e3 | /documentation/faq/qdrant-fundamentals/ | Qdrant Fundamentals | Vectors > Vectors in the collection don't match what I uploaded. | documentation | false | Qdrant Fundamentals > Vectors > Vectors in the collection don't match what I uploaded.
There are two possible reasons for this:
- You used the `Cosine` distance metric in the [collection settings](/documentation/manage-data/collections/#collections). In this case, Qdrant pre-normalizes your vectors for faster distanc... | [
0.039845313876867294,
-0.05316721647977829,
0.004484651144593954,
-0.027826255187392235,
0.07455659657716751,
-0.027373114600777626,
0.05133246257901192,
0.06097711995244026,
0.08568639308214188,
-0.050824496895074844,
0.028324397280812263,
0.020243212580680847,
0.03980899602174759,
-0.003... |
cdfbc12d-fb7c-5cab-880c-61e8e5ae5700 | /documentation/platforms/n8n/ | N8N | Prerequisites | documentation | false | N8N > Prerequisites
1. A Qdrant instance to connect to. You can get a free cloud instance at [cloud.qdrant.io](https://cloud.qdrant.io/).
2. A running N8N instance. You can learn more about using the N8N cloud or self-hosting [here](https://docs.n8n.io/choose-n8n/). | [
-0.009072279557585716,
-0.03505507484078407,
-0.032881710678339005,
-0.048145707696676254,
0.04071984812617302,
-0.04300909489393234,
0.018180295825004578,
0.025041747838258743,
0.06173025071620941,
-0.009714765474200249,
0.04386261850595474,
0.012646298855543137,
0.04043234512209892,
-0.0... |
ce30a44b-0532-5e0c-9b83-0ac819ae05db | /documentation/production-operations/vector-search-in-production/ | Vector Search in Production: Scaling, HA & Tuning Guide | 5. Tips for Proper Database Administration > Security & Governance | documentation | true | Vector Search in Production: Scaling, HA & Tuning Guide > 5. Tips for Proper Database Administration > Security & Governance
**Use Case:** A manufacturing company created a "Super Chatbot" that relied on many organizational components. The company needed to ensure secure communication between their application compone... | [
0.009203949011862278,
-0.06684673577547073,
-0.08526460826396942,
-0.022282570600509644,
0.005844018422067165,
-0.007932907901704311,
0.04613348841667175,
0.029740365222096443,
0.06313040852546692,
0.011839187704026699,
0.0011973702348768711,
-0.015791799873113632,
0.0664607509970665,
-0.0... |
ce650ca0-2e0f-5c6d-8ffa-85239609c879 | /documentation/search/text-search/text-filtering/ | Text Filtering | Full-Text Filtering > Text Processing | documentation | true | Text Filtering > Full-Text Filtering > Text Processing
These text processing steps can be configured when creating a [full-text index](/documentation/manage-data/indexing/#full-text-index). For example, to create a text index on the `title` field with ASCII folding enabled:
```python
from qdrant_client import QdrantC... | [
0.03607138991355896,
-0.061935748904943466,
-0.028045833110809326,
-0.03926229104399681,
0.0511089488863945,
-0.03677184507250786,
0.03110319934785366,
0.041335638612508774,
0.052710406482219696,
-0.009824358858168125,
0.07349500060081482,
0.014409977942705154,
0.06964439153671265,
-0.0458... |
cee92b08-6d78-5d68-b9bb-90b0d8bedcac | /documentation/faq/qdrant-fundamentals/ | Qdrant Fundamentals | Compatibility > If I need to add a payload index after the HNSW index has already been built, how do I trigger a full reindex? | documentation | false | Qdrant Fundamentals > Compatibility > If I need to add a payload index after the HNSW index has already been built, how do I trigger a full reindex?
Changing `m` or `ef_construct` automatically triggers a full background HNSW rebuild. For cases where only a payload index is being added, make a minimal change to `ef_co... | [
0.01580544374883175,
-0.019320186227560043,
-0.032408349215984344,
-0.046295229345560074,
0.04460001364350319,
-0.0432962067425251,
0.03036470152437687,
0.045668136328458786,
0.06652083247900009,
-0.045365601778030396,
0.05207963287830353,
0.03671570494771004,
0.029283827170729637,
-0.0539... |
ceebeb32-22d9-53aa-80e1-a63300ce109f | /documentation/edge/edge-fastembed-embeddings/ | On-Device Embeddings | documentation | true | On-Device Embeddings
When using Python, you can use the [FastEmbed](/documentation/fastembed/) library to generate embeddings for use with Qdrant Edge. FastEmbed provides multimodal models that run efficiently on edge devices to generate vector embeddings from text and images.
To generate sparse BM25 embeddings for k... | [
0.06093544512987137,
-0.03704521432518959,
-0.029619211331009865,
-0.013711935840547085,
0.07203076034784317,
-0.047326814383268356,
0.08924809843301773,
0.047496579587459564,
0.03168482333421707,
-0.01634969376027584,
0.05537230893969536,
0.02337588556110859,
0.01162561122328043,
-0.05203... | |
cef3fbc3-db3b-5117-ac28-e210ec8f5f30 | /documentation/production-operations/predictable-search-latency/ | Configure Qdrant's Optimizer for Predictable Search Latency | prevent_unoptimized Buys Speed | documentation | false | Configure Qdrant's Optimizer for Predictable Search Latency > prevent_unoptimized Buys Speed
With continuous indexing, the experimental `prevent_unoptimized` flag (available since Qdrant 1.17.1) can reduce query latency under heavy write load. It works on the write path: once a growing segment's data crosses `indexing... | [
0.02317267842590809,
-0.0437680222094059,
-0.054261215031147,
-0.01636895351111889,
0.045353613793849945,
-0.04819266125559807,
0.028578322380781174,
0.006294040475040674,
0.08850323408842087,
-0.030042467638850212,
0.024050455540418625,
-0.009950589388608932,
0.06584620475769043,
-0.03251... |
cf2ad797-87ae-57af-a671-d593a4c76d2e | /documentation/tutorials-operations/gpu-accelerated-hnsw-indexing/ | GPU-Accelerated HNSW Indexing | Monitoring Indexing and Query Latency > Query While Indexing | documentation | true | GPU-Accelerated HNSW Indexing > Monitoring Indexing and Query Latency > Query While Indexing
Run two coroutines concurrently:
- One polls `GET /collections/{collection}/optimizations` every 0.2s until every running or queued optimization has finished, recording each snapshot.
- The other repeatedly queries the collect... | [
0.014196932315826416,
-0.04078676179051399,
-0.03338124230504036,
-0.020932678133249283,
0.047651395201683044,
-0.04305974021553993,
0.056726641952991486,
0.02694668434560299,
0.09229162335395813,
-0.018803812563419342,
0.07236844301223755,
0.009273279458284378,
0.06398747861385345,
-0.005... |
cf341ba8-5a2d-5142-b1c1-fbedd39f5da8 | /documentation/snapshots/ | Snapshots | Snapshots for the whole storage | documentation | true | Snapshots > Snapshots for the whole storage
*Available as of v0.8.5*
Sometimes it might be handy to create snapshot not just for a single collection, but for the whole storage, including collection aliases.
Qdrant provides a dedicated API for that as well. It is similar to collection-level snapshots, but does not req... | [
-0.004714122042059898,
-0.02790716104209423,
-0.016989273950457573,
-0.05558687821030617,
0.02501475252211094,
-0.01808626390993595,
0.04067675396800041,
0.019342120736837387,
0.08763472735881805,
-0.051234569400548935,
0.06479815393686295,
0.04598189517855644,
0.05319395661354065,
-0.0171... |
cf37f222-6faf-5b56-97f3-7ef365a1a0c6 | /course/essentials/day-5/universal-query-api/ | The Universal Query API | Dense Retrieval + ColBERT Rerank | course | true | The Universal Query API > Dense Retrieval + ColBERT Rerank
While the previous lesson showed how to use ColBERT directly for retrieval, **in production systems reranking is the more common pattern**. The reason is practical: ColBERT's brute-force MaxSim scoring on an entire large collection can be slow. Instead, you co... | [
0.029310395941138268,
-0.019365400075912476,
-0.04352480173110962,
-0.03534728288650513,
0.07488523423671722,
-0.04409052059054375,
0.033520154654979706,
0.04346694052219391,
0.0878848060965538,
-0.017459126189351082,
0.049638692289590836,
0.04041161388158798,
0.07098853588104248,
-0.01281... |
cf5a9129-c6f0-5b88-829f-dc43136d53e7 | /documentation/scaling/resilience/ | Resilience | Setting Up a Resilient Qdrant Cluster | documentation | false | Resilience > Setting Up a Resilient Qdrant Cluster
Two factors determine the resilience of a Qdrant cluster:
- **[Replication factor](/documentation/scaling/distributed_deployment/#replication)** controls how many copies of each shard exist. More copies mean your data survives the loss of more nodes.
- **[Node count]... | [
0.03849688917398453,
-0.02898542210459709,
-0.05085131525993347,
-0.042915090918540955,
0.08105809986591339,
-0.024381298571825027,
0.044062916189432144,
0.010650296695530415,
0.050660617649555206,
-0.04477233439683914,
0.05017808452248573,
0.002046701731160283,
0.06499715149402618,
-0.024... |
cf65cbe1-261f-5cf1-b873-8b52e7eaa858 | /documentation/cloud/cluster-scaling/ | Scale Clusters | Resharding | documentation | true | Scale Clusters > Resharding
*Available as of Qdrant v1.13.0*
Resharding is exclusively available on multi-node clusters across our cloud offering, including Hybrid and Private Cloud.
When creating a collection, it has a specific number of shards. The ideal number of shards might change as your cluster evolves.
Resh... | [
0.012032641097903252,
-0.03253636136651039,
-0.028573816642165184,
-0.04370313882827759,
0.08349055796861649,
-0.05571279674768448,
0.03925681114196777,
0.04239599034190178,
0.06241927295923233,
-0.0801922008395195,
0.024334048852324486,
0.011565872468054295,
0.024736978113651276,
-0.01961... |
cf766e43-554a-5cd3-9a7b-3d70d4f080cc | /documentation/tutorials-operations/secure-qdrant/ | Secure a Self-Hosted Qdrant Instance | Step 4: Enable a Read-Only API Key | documentation | true | Secure a Self-Hosted Qdrant Instance > Step 4: Enable a Read-Only API Key
Issue a separate [read-only API key](/documentation/security/#read-only-api-key) for services that only need to read data. With this key, a client application can search and read but cannot upsert, delete, or modify data.
Set the `QDRANT__SERVI... | [
0.05090835317969322,
-0.045772090554237366,
-0.04556075111031532,
-0.0398319847881794,
0.043224118649959564,
-0.0367475189268589,
0.04072188213467598,
0.039143189787864685,
0.05132488161325455,
-0.007498110644519329,
0.016190923750400543,
0.025219593197107315,
0.08142950385808945,
-0.04069... |
cf795d74-2574-5b33-a717-b22b472668cf | /course/essentials/day-1/embedding-models/ | Points, Vectors and Payloads | Vector Types in Qdrant > Dense Vectors | course | false | Points, Vectors and Payloads > Vector Types in Qdrant > Dense Vectors
At the core of every vector is a set of numbers, which together form a representation of the data in a multi-dimensional space.
Dense vectors are the typical vector representation used in vector search, generated by the majority of the embedding mo... | [
0.04237750172615051,
-0.04643529653549194,
-0.052195608615875244,
-0.040072228759527206,
0.03880167752504349,
-0.043478723615407944,
0.09289176762104034,
0.08226245641708374,
0.09740528464317322,
-0.048359889537096024,
0.04058269411325455,
0.04287228733301163,
0.04840666428208351,
-0.00201... |
cf8323f2-b98e-5c78-a229-99f2f40ec866 | /course/essentials/ | Qdrant Essentials Course | How the course works | course | false | Qdrant Essentials Course > How the course works
- icon: /icons/outline/training-purple.svg
title: Video-first lessons
content: Clear, concise modules by the Qdrant team
- icon: /icons/outline/hacker-purple.svg
title: Final project
content: Ship a production-ready vector search app
- icon: /icons/outline/simila... | [
0.032615020871162415,
-0.0516558475792408,
-0.06476680189371109,
-0.013783388771116734,
0.04016526788473129,
-0.0017480128444731236,
0.05249300226569176,
0.0484723299741745,
0.04836137592792511,
0.006210979074239731,
0.007187127601355314,
0.06218934804201126,
0.03533954918384552,
-0.051046... |
cf8c613a-f9a7-5eb7-b6d4-fa1379bb2a99 | /articles/minicoil-en-es/ | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier | Results > Cross-Lingual Retrieval | article | false | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier > Results > Cross-Lingual Retrieval
| Direction | System | MRR@10 | nDCG@10 | R@100 |
| --- | --- | --- | --- | --- |
| EN→ES | translate-then-BM25 | **0.730** | **0.757** | 0.941 |
| EN→ES | miniCOIL EN-ES | 0.711 | 0.743 | **0.963** |
| ES→EN | tran... | [
0.03884474188089371,
-0.04581763222813606,
-0.03421753644943237,
-0.03310054540634155,
0.05411352217197418,
-0.05138704553246498,
0.06572100520133972,
0.040504369884729385,
0.06547025591135025,
-0.033754076808691025,
0.010413370095193386,
0.005983167793601751,
0.052369676530361176,
-0.0152... |
cf8d08e8-f01f-54c6-a871-5da9a7c71172 | /course/essentials/day-1/distance-metrics/ | Distance Metrics | Summary: When to Use Each Metric | course | false | Distance Metrics > Summary: When to Use Each Metric
If you are training your own model or designing custom features, use these guidelines:
| Metric | Common Application | Why? |
| :--- | :--- | :--- |
| **Cosine Similarity** | NLP, Semantic Search | Ignores magnitude; focuses on semantic meaning (direction). |
| **Do... | [
0.0168939009308815,
-0.021804098039865494,
-0.051733385771512985,
-0.019229482859373093,
0.07072056829929352,
-0.0205918587744236,
0.03713449835777283,
0.0223068930208683,
0.08856912702322006,
-0.02252201922237873,
0.04051612317562103,
-0.018221478909254074,
0.06358864158391953,
-0.0191235... |
cf951d11-769a-5526-9f8e-4513d9b4cefc | /documentation/cloud-rbac/permission-reference/ | Permission Reference | Identity and Access Management | documentation | false | Permission Reference > Identity and Access Management
Permissions for users, user roles, management keys, and invitations.
| Permission | Description |
|------------|------------|
| `read:roles` | View roles in the Access Management page. |
| `write:roles` | Create and modify roles in the Access Management page. |
| ... | [
0.01968899555504322,
-0.003404526039958,
-0.058728981763124466,
-0.029792774468660355,
0.04664033651351929,
-0.014670420438051224,
0.0169097688049078,
0.10149137675762177,
0.07159864902496338,
-0.016774723306298256,
0.02918611839413643,
0.0025895126163959503,
0.05211932584643364,
-0.027739... |
cfb3c59b-c976-5d53-aecb-2158c8828d9b | /course/multi-vector-search/module-1/multi-vector-in-qdrant/ | Multi-Vector Embeddings in Qdrant | What's Next | course | false | Multi-Vector Embeddings in Qdrant > What's Next
You've learned how to configure Qdrant for multi-vector search, from creating collections with MaxSim comparators to indexing documents and executing queries. The key takeaways:
- **Multi-vector collections** require specific configuration: `multivector_config` with `MA... | [
0.05012216418981552,
-0.04417935758829117,
-0.03200154006481171,
-0.03169643133878708,
0.07845510542392731,
-0.042407725006341934,
0.053594108670949936,
0.05835970118641853,
0.05313391610980034,
-0.032999783754348755,
0.025875741615891457,
0.024786734953522682,
0.05221043527126312,
-0.0316... |
cffc3925-bb94-59ae-bd5f-841fa7658e37 | /documentation/tutorials-search-engineering/pdf-retrieval-at-scale/ | Multivector Document Retrieval | Setup | documentation | true | Multivector Document Retrieval > Setup
Install & import required libraries
```python
# pip install "colpali_engine>=0.3.1" "qdrant-client>=1.12.0"
import torch
from colpali_engine.models import ColPali, ColPaliProcessor
from qdrant_client import QdrantClient, models
```
To run these experiments, we’re using a **Qdr... | [
0.010999089106917381,
-0.05535326525568962,
-0.03251771256327629,
-0.0071050142869353294,
0.08084453642368317,
-0.02209097146987915,
0.056749872863292694,
0.047675326466560364,
0.02306167408823967,
-0.012555061839520931,
0.0015282431850209832,
0.015288107097148895,
0.06057422608137131,
-0.... |
d00f3592-b2ef-5269-bcb7-9e32e2e8384b | /documentation/production-operations/bulk-data-import/ | Bulk Uploading Data to Qdrant | Why Bulk Uploading Matters | documentation | false | Bulk Uploading Data to Qdrant > Why Bulk Uploading Matters

When you start using Qdrant at scale, one of the first challenges you may run into is uploading large amounts of dat... | [
0.006814508233219385,
-0.0368364080786705,
-0.04274153709411621,
-0.04703991487622261,
0.0814281553030014,
-0.0554502010345459,
0.06900553405284882,
0.055994994938373566,
0.09570761024951935,
-0.05981062725186348,
0.00887331087142229,
0.038739562034606934,
0.024811143055558205,
-0.05525494... |
d01d4b05-867c-5d37-8c1a-f93e9a97ac91 | /documentation/frameworks/llama-index/ | LlamaIndex | documentation | true | LlamaIndex
Llama Index acts as an interface between your external data and Large Language Models. So you can bring your
private data and augment LLMs with it. LlamaIndex simplifies data ingestion and indexing, integrating Qdrant as a vector index.
Installing Llama Index is straightforward if we use pip as a package ... | [
0.02538939192891121,
-0.06301717460155487,
-0.06581389904022217,
-0.04635564610362053,
0.0674990713596344,
-0.02462225779891014,
0.08761611580848694,
0.04762185364961624,
0.06798862665891647,
-0.03465138003230095,
0.015298079699277878,
0.03266708180308342,
0.012886243872344494,
-0.02797570... | |
d02f5105-6b05-5d4f-9991-d977bc92062e | /articles/relevance-feedback/ | Relevance Feedback in Qdrant | Experiments | article | false | Relevance Feedback in Qdrant > Experiments
Before building a new interface, we needed to confirm our instinct: that a relevance feedback scoring can surface relevant documents that initial retrieval misses. | [
0.013448982499539852,
-0.025022216141223907,
-0.030393395572900772,
-0.07548043131828308,
0.03341865912079811,
-0.021727105602622032,
0.055374082177877426,
0.05030630901455879,
0.060821328312158585,
-0.02467077225446701,
0.017431719228625298,
0.057945989072322845,
0.10913744568824768,
-0.0... |
d0339a53-5f7c-58ea-8475-f8e0dbe4b4a0 | /articles/what-is-a-vector-database/ | What is a Vector Database? | The Core Functionalities of Vector Databases > 1. Indexing | article | false | What is a Vector Database? > The Core Functionalities of Vector Databases > 1. Indexing
Indexing your vectors is like creating an entry in a traditional database. But for vector databases, this step is very important. Vectors need to be indexed in a way that makes them easy to search later on.
#### 1.1 HNSW Indexing
... | [
0.014491444453597069,
-0.04715360328555107,
-0.0733390673995018,
-0.027150504291057587,
0.04690202698111534,
-0.03999089077115059,
0.06199062243103981,
0.04349811375141144,
0.08204025775194168,
-0.038145534694194794,
0.034636933356523514,
0.02619074657559395,
0.04420953616499901,
-0.006506... |
d034f283-1321-5fa4-a175-b17377bea70a | /documentation/examples/graphrag-qdrant-neo4j/ | GraphRAG with Qdrant and Neo4j | Implementation > Retrieval & Generation > Building a Retriever | documentation | true | GraphRAG with Qdrant and Neo4j > Implementation > Retrieval & Generation > Building a Retriever
The retriever integrates vector search and graph data, enabling semantic similarity searches with Qdrant and fetching relevant graph data from Neo4j. This enriches the RAG process and allows for more informed responses.
``... | [
0.029271885752677917,
-0.052553147077560425,
-0.07553432136774063,
-0.039031561464071274,
0.07494612038135529,
-0.01973695121705532,
0.05217931047081947,
0.05168319493532181,
0.0378524512052536,
-0.040473490953445435,
0.03224590793251991,
0.02170184627175331,
0.06152256950736046,
-0.018612... |
d0421a28-6fda-50f2-8cb6-4050a38a3472 | /course/beginners/module-0/qdrant-cloud/ | Qdrant Setup | Common Issues | course | false | Qdrant Setup > Common Issues
- **Authentication error:** Recheck the API key and the `api-key` header. A stray space or a missing character is the usual cause.
- **Connection error:** Confirm the cluster is running and the region URL is correct. Some workplace networks block outbound connections, so try from a persona... | [
0.026552699506282806,
-0.013437016867101192,
-0.02830485627055168,
-0.02166428230702877,
0.07173934578895569,
-0.04512247070670128,
0.02881084568798542,
0.046942759305238724,
0.053854554891586304,
-0.03850902244448662,
0.009798022918403149,
0.002140830270946026,
0.05265822261571884,
-0.055... |
d05ba207-f6d5-523e-ad85-ebf39b3e22e7 | /articles/minicoil/ | miniCOIL: on the Road to Usable Sparse Neural Retrieval | Eyes on the Prize: Usable Sparse Neural Retrieval > Inspired by COIL | article | false | miniCOIL: on the Road to Usable Sparse Neural Retrieval > Eyes on the Prize: Usable Sparse Neural Retrieval > Inspired by COIL
1. We should **abandon end-to-end training on a relevance objective** to get a model performant on out-of-domain data. There is not enough data to train a model able to generalize.
2. We shoul... | [
0.06659243255853653,
-0.03583914041519165,
-0.014685211703181267,
-0.017281830310821533,
0.06028103455901146,
-0.07208137214183807,
0.035716693848371506,
0.003233758732676506,
0.05139775946736336,
-0.016573261469602585,
0.029409468173980713,
0.05790218710899353,
0.09917137026786804,
-0.028... |
d08f8f6b-c918-5dfe-b23c-d658590bfa4f | /documentation/edge/edge-api/updating-data/ | Updating Data | Update Operations > Parameters | documentation | false | Updating Data > Update Operations > Parameters
| Variant | Covers |
| --- | --- |
| `PointOperation` | Upserting, deleting, and syncing points. |
| `VectorOperation` | Updating and deleting named vectors on existing points. |
| `PayloadOperation` | Setting, overwriting, deleting, and clearing payload. |
| `FieldIndexO... | [
0.041136085987091064,
-0.05420323461294174,
-0.03230779618024826,
-0.03390473872423172,
0.06525517255067825,
0.006554238963872194,
0.008163802325725555,
0.019795460626482964,
0.04823782667517662,
-0.0247975904494524,
0.04949764162302017,
0.04020899534225464,
0.025225646793842316,
-0.025766... |
d0909c34-17d8-5198-9852-d3279b0e7592 | /documentation/private-cloud/changelog/ | Changelog | 1.9.15 (2026-09-04) | documentation | false | Changelog > 1.9.15 (2026-09-04)
| Component | Version |
|-------------------------|---------|
| qdrant-kubernetes-api | v1.37.7 |
| operator | 2.34.9 |
| qdrant-cluster-manager | v0.3.22 |
| qdrant-cluster-exporter | 1.8.0 |
Latest validated Qdrant version: 1.19.0
* Fixed an edge c... | [
0.0457775741815567,
-0.023442918434739113,
-0.05789245665073395,
-0.03754467889666557,
0.05405271798372269,
-0.04828401654958725,
0.07582572847604752,
0.023992309346795082,
0.045734431594610214,
-0.04357299208641052,
0.05123858526349068,
0.03020365908741951,
0.04752057045698166,
-0.0578150... |
d090d799-f396-5de9-ac91-ab58e2306b21 | /course/multi-vector-search/module-3/evaluating-pipelines/ | Evaluating Search Pipelines | Evaluation with ranx | course | true | Evaluating Search Pipelines > Evaluation with ranx
The **ranx** library provides battle-tested implementations of IR metrics. Here's how to compare all our pipelines:
```python
from ranx import Qrels, Run, compare
import time
# Define ground truth - which documents are relevant to which queries
qrels_dict = {
"c... | [
0.03304142504930496,
-0.014337016269564629,
-0.03690878301858902,
-0.03570706769824028,
0.06670936942100525,
-0.011962817050516605,
0.0861469954252243,
0.014163484796881676,
0.018190519884228706,
-0.01870040036737919,
0.050292838364839554,
0.017127247527241707,
0.05217697098851204,
-0.0075... |
d0baee08-06fb-50e2-90c9-3871284efc1d | /documentation/embeddings/ollama/ | Ollama | Models | documentation | false | Ollama > Models
Ollama lists its embedding models at [ollama.com/search?c=embedding](https://ollama.com/search?c=embedding). These are good starting points:
| Model | Dimensions | Max Tokens | Download Size |
|---|---|---|---|
| `qwen3-embedding:0.6b` | 1024 (reducible to 32) | 32,768 | 639 MB |
| `qwen3-embedding:4b... | [
0.029308712109923363,
-0.027856916189193726,
-0.038784291595220566,
-0.06942916661500931,
0.0869799479842186,
-0.0634123906493187,
0.05732252448797226,
0.08647283166646957,
0.039797622710466385,
-0.04007449001073837,
0.022693216800689697,
0.01452257763594389,
0.02391972206532955,
-0.044297... |
d0bf1c16-6430-5014-b873-de515e240199 | /documentation/migration-guidance/pre-migration-baseline/ | Pre-Migration Baseline | What to Capture | documentation | false | Pre-Migration Baseline > What to Capture
There are four pieces of information that need to be accounted for when establishing a baseline: collection/index inventory, metadata samples, baseline search results, and system configuration snapshots. | [
-0.004218224436044693,
-0.03878726810216904,
-0.015976866707205772,
-0.0723671168088913,
0.018270153552293777,
-0.04709604009985924,
0.03346534073352814,
0.09456734359264374,
0.09137538820505142,
-0.0177554190158844,
0.01566084660589695,
0.023867743089795113,
0.046053942292928696,
-0.04977... |
d0c675d1-de34-5efb-8cfa-debd9505a587 | /documentation/manage-data/points/ | Points | Point IDs | documentation | true | Points > Point IDs
Qdrant supports using both `64-bit unsigned integers` and `UUID` as identifiers for points.
Examples of UUID string representations:
- simple: `936DA01F9ABD4d9d80C702AF85C822A8`
- hyphenated: `550e8400-e29b-41d4-a716-446655440000`
- urn: `urn:uuid:F9168C5E-CEB2-4faa-B6BF-329BF39FA1E4`
That means ... | [
0.054507359862327576,
-0.01796221360564232,
-0.03176545351743698,
-0.02471151389181614,
0.05484197288751602,
-0.03868430107831955,
0.03219485282897949,
0.0550786629319191,
0.09128732234239578,
-0.0006920204032212496,
0.028065010905265808,
0.013017088174819946,
0.030598795041441917,
-0.0020... |
d0f98e5b-5d80-5db3-8321-9865bc229755 | /documentation/search/search-relevance/ | Search Relevance | Score Boosting > Decay Functions | documentation | false | Search Relevance > Score Boosting > Decay Functions
Decay functions enable you to modify the score based on how far a value is from a target using a linear, exponential, or Gaussian decay function. For all decay functions, these are the available parameters:
| Parameter | Default | Description ... | [
0.01913568191230297,
0.009112021885812283,
-0.041004009544849396,
-0.0004729164356831461,
0.03815768286585808,
-0.01792510226368904,
0.008387180976569653,
0.06882171332836151,
0.10056779533624649,
-0.01749391108751297,
0.028968289494514465,
-0.007723760325461626,
0.015549854375422001,
0.00... |
d102b855-a4d0-5dee-9a8a-35317fcd9f43 | /documentation/scaling/ | Scaling & Resilience | Vertical vs. Horizontal Scaling | documentation | false | Scaling & Resilience > Vertical vs. Horizontal Scaling
Vertical scaling means adding more CPU, RAM, or disk to an existing node. A single node can typically hold up to about 100 million vectors, depending on dimensionality and quantization. RAM usage approaching 80% is the main signal that it's time to resize. See [Ve... | [
0.042976994067430496,
-0.011787486262619495,
-0.055218711495399475,
-0.04405095800757408,
0.08382843434810638,
-0.013411544263362885,
0.0887548103928566,
0.05028919130563736,
0.09291382879018784,
-0.03612592816352844,
0.02579013630747795,
-0.00941804051399231,
0.029297739267349243,
-0.0132... |
d1589f03-020a-5b43-8c44-ab226c3ed02b | /course/beginners/ | Qdrant Beginner Course | course | false | Qdrant Beginner Course
**Learn the fundamentals of vector search**
Understand why traditional search struggles and how modern semantic search improves it. Learn about embeddings, distance metrics, and hybrid search systems.
- icon: /icons/outline/play-white.svg
title: 6 modules
content: From setting up dependenc... | [
0.03223497420549393,
-0.059625864028930664,
-0.026474574580788612,
-0.0246925987303257,
0.030213147401809692,
-0.027240384370088577,
0.05987045168876648,
0.07111389189958572,
0.06829189509153366,
-0.0011960658011958003,
0.0008561561116948724,
0.016255313530564308,
0.026175806298851967,
-0.... | |
d19a3141-ab0a-5bb1-843e-a1fc73524b82 | /documentation/cloud-rbac/permission-reference/ | Permission Reference | Cluster > Cluster Data | documentation | false | Permission Reference > Cluster > Cluster Data
| Permission | Description |
|------------|------------|
| `read:cluster_data` | View cluster data, used for the Cluster UI button on Cluster Details. [Maps to global `read-only` JWT access for the cluster.](/documentation/security/) |
| `write:cluster_data` | View and mod... | [
0.05016862228512764,
0.0034533196594566107,
-0.05764679238200188,
-0.05234498530626297,
0.05779629945755005,
0.0032021631486713886,
0.02025609090924263,
0.06441889703273773,
0.040682315826416016,
-0.02090379036962986,
0.032156191766262054,
0.01928618922829628,
0.08318676054477692,
-0.01867... |
d1bf06cb-1a06-52ba-afd7-b3773e3252ea | /documentation/production-operations/memory-tiers/ | Memory Tiers in Qdrant: What to Use and When | Use The Cold Tier With Quantization | documentation | false | Memory Tiers in Qdrant: What to Use and When > Use The Cold Tier With Quantization
If disk footprint matters more than just raw speed, consider adding quantization before fully committing to the cold tier. Every **first touch on a cold, unquantized structure comes from disk**, and a single HNSW traversal touches enoug... | [
0.04778258875012398,
-0.019442634657025337,
-0.06958890706300735,
-0.05539412796497345,
0.0632469579577446,
-0.02071377821266651,
0.032981861382722855,
0.009681524708867073,
0.1070060059428215,
0.005392895545810461,
-0.0009173118160106242,
-0.016878098249435425,
0.06784140318632126,
-0.024... |
d1c59d44-e59e-56fd-8859-6ca7cb307ae5 | /documentation/search-tuning/before-tuning-a-qdrant-collection/ | What to Check Before Tuning a Qdrant Collection | Start with One Change | documentation | false | What to Check Before Tuning a Qdrant Collection > Start with One Change
Record the current relevance metric and p95 latency for a representative query set. Choose one low-cost change from the symptom table, validate it on fresh queries, and keep it only if the gain survives. Once you have that baseline, [Candidate Dep... | [
0.03336336463689804,
-0.073853999376297,
-0.019090263172984123,
-0.037627581506967545,
0.01637927256524563,
-0.04836549609899521,
0.04096413776278496,
0.02103330008685589,
0.09884804487228394,
-0.027322372421622276,
0.04753294214606285,
0.03717999532818794,
0.08462391048669815,
-0.04300118... |
d1c644dd-0704-52c1-a990-766be38f7c07 | /documentation/data-synchronization/with-postgres/ | Keeping Postgres and Qdrant in Sync | Comparison Matrix | documentation | false | Keeping Postgres and Qdrant in Sync > Comparison Matrix
| Dimension | Tier 1: Dual-Write | Tier 2: Outbox | Tier 3: CDC |
| :---- | :---- | :---- | :---- |
| **Sync complexity** | Low | Medium | High |
| **Extra infrastructure** | None | Outbox table + worker | Redpanda + Debezium + consumer |
| **Consistency model** ... | [
0.026352891698479652,
-0.010026040486991405,
-0.042880382388830185,
-0.01715446263551712,
0.044204164296388626,
-0.056042809039354324,
0.09276468306779861,
0.00524451769888401,
0.07097198069095612,
-0.026416772976517677,
0.05285751447081566,
-0.00009931034583132714,
0.030201559886336327,
-... |
d1cdca00-b58c-5c9a-ade5-38c7c7fee6e4 | /articles/minicoil/ | miniCOIL: on the Road to Usable Sparse Neural Retrieval | Eyes on the Prize: Usable Sparse Neural Retrieval > Bag-of-words in 4D | article | false | miniCOIL: on the Road to Usable Sparse Neural Retrieval > Eyes on the Prize: Usable Sparse Neural Retrieval > Bag-of-words in 4D
COIL uses 32 values to describe one term. Do we need this many? How many words with 32 separate meanings could we name without additional research?
Yet, even if we use fewer values in COIL ... | [
0.05714455991983414,
-0.03748898580670357,
-0.027874218299984932,
-0.026750244200229645,
0.07469838112592697,
-0.049251969903707504,
0.023229995742440224,
0.040684934705495834,
0.06256541609764099,
-0.01723652519285679,
0.028070108965039253,
0.035511910915374756,
0.0726456493139267,
-0.027... |
d1df090d-7387-519d-96c7-4b78b5e138c9 | /documentation/ops-configuration/memory-tiers/ | Memory Tiers | Optimizing for Disk-Based Retrieval > Async I/O | documentation | true | Memory Tiers > Optimizing for Disk-Based Retrieval > Async I/O
Async I/O lets Qdrant issue disk reads concurrently instead of one at a time, reducing how long a query waits on disk when structures are `cold`. This uses [io_uring](/articles/io_uring/), a Linux kernel interface for asynchronous I/O, and requires a kerne... | [
0.04196688532829285,
-0.03885351121425629,
-0.03528521582484245,
-0.009522751905024052,
0.08021719008684158,
-0.011458199471235275,
0.013459163717925549,
0.008277821354568005,
0.07373783737421036,
-0.0003953208215534687,
-0.006076753605157137,
-0.014265964739024639,
0.07217202335596085,
-0... |
d20223b9-9aac-5e51-927f-65b6284804ed | /documentation/search/search/ | Search | Grouping API > Lookup in Groups | documentation | true | Search > Grouping API > Lookup in Groups
When the points in a group share large fields like titles, abstracts, or full document vectors, copying that data onto every point inflates storage and forces you to rewrite every chunk whenever a shared field changes.
`with_lookup` solves this. Store the shared data once in a... | [
0.001169805764220655,
-0.03951197862625122,
-0.045913875102996826,
-0.017484793439507484,
0.07765445858240128,
-0.034935664385557175,
0.05102916434407234,
0.007885511964559555,
0.04537005349993706,
-0.022405216470360756,
0.03499311953783035,
0.01478708628565073,
0.0062434072606265545,
-0.0... |
d2037ebe-8cac-5943-8902-88d00c99a1fd | /documentation/manage-data/multitenancy/ | Multitenancy | Tiered Multitenancy > Write to a Tiered Multitenant Collection | documentation | true | Multitenancy > Tiered Multitenancy > Write to a Tiered Multitenant Collection
Now you can start uploading data into the collection. Specify a **shard key selector** in each request to reach the correct shard. The shard key selector needs to specify two keys:
- `target` shard - name of the tenant's dedicated shard (wh... | [
-0.00856828037649393,
-0.013403987511992455,
-0.056953221559524536,
0.001102335867471993,
0.04911785572767258,
-0.0441308468580246,
0.06458008289337158,
0.04689934849739075,
0.05146877095103264,
-0.016504663974046707,
0.0036622968036681414,
0.046694256365299225,
0.03473968058824539,
-0.024... |
d24d32ed-8f03-59a1-a514-52df959f520b | /documentation/tutorials-build-essentials/agentic-rag-langgraph/ | Agentic RAG with LangGraph | Implementation > Document Processing | documentation | true | Agentic RAG with LangGraph > Implementation > Document Processing
Before we can create our agent, we need to process and store the documentation. We’ll be working with two datasets from Hugging Face: their general documentation and Transformers-specific documentation.
Here’s our document preprocessing function:
```p... | [
0.006295792292803526,
-0.06671162694692612,
-0.046266742050647736,
-0.023165538907051086,
0.08016899973154068,
-0.05848013609647751,
0.026668453589081764,
0.026116259396076202,
0.040685709565877914,
-0.06006829813122749,
0.003272054484114051,
0.049024201929569244,
0.054037973284721375,
-0.... |
d2569fa7-1fa8-5666-b6fb-26cce3976e0b | /documentation/edge/edge-data-synchronization-patterns/ | Data Synchronization Patterns | Update Qdrant Edge with Server-Side Changes | documentation | true | Data Synchronization Patterns > Update Qdrant Edge with Server-Side Changes
To keep an Edge Shard updated with new data from a server collection, you can periodically download and apply a snapshot. Restoring a full snapshot every time would create unnecessary overhead. Instead, you can use partial snapshots to restore... | [
0.004343545064330101,
-0.023368408903479576,
-0.032637547701597214,
-0.030504407361149788,
0.01824302226305008,
-0.03354423865675926,
0.07026828825473785,
-0.0007955119363032281,
0.07949469238519669,
-0.030288100242614746,
0.038285598158836365,
0.056723907589912415,
0.044507622718811035,
-... |
d26dc8bd-a794-537e-8362-ba507c14b46a | /documentation/snapshots/ | Snapshots | Retrieve snapshot | documentation | true | Snapshots > Retrieve snapshot
Only available through the REST API for the time being.
To download a specified snapshot from a collection as a file:
```http
GET /collections/{collection_name}/snapshots/{snapshot_name}
``` | [
-0.009668339043855667,
-0.053408797830343246,
0.0021426971070468426,
-0.08884868770837784,
0.019386760890483856,
-0.03721453621983528,
0.005787927191704512,
0.018907252699136734,
0.055753983557224274,
-0.040313154458999634,
0.05257885530591011,
0.05047018453478813,
0.06321783363819122,
-0.... |
d2737229-f90b-5ea9-93d0-3fdcc49075f8 | /documentation/cloud/cluster-monitoring/ | Monitor Clusters | Grafana Dashboard > Cluster System Metrics /sys_metrics | documentation | false | Monitor Clusters > Grafana Dashboard > Cluster System Metrics /sys_metrics
| Name | Type | Meaning |
|-----------------------------------... | [
0.014051228761672974,
-0.028813393786549568,
-0.04303719475865364,
0.0024002857971936464,
0.046211350709199905,
-0.07127334922552109,
0.05898235738277435,
0.05410290136933327,
0.0590745285153389,
-0.046843890100717545,
0.03263350948691368,
0.015719719231128693,
0.028067775070667267,
-0.008... |
d274ca93-6300-5eff-8d27-a73e1ea290dd | /course/beginners/module-3/hybrid-search-in-qdrant/ | Setting Up Hybrid Search in Qdrant | Step 3: Hybrid Query with Fusion | course | true | Setting Up Hybrid Search in Qdrant > Step 3: Hybrid Query with Fusion
```python
def hybrid_search(query_text, limit=4):
return client.query_points(
collection_name="products",
# One sub-query per vector, both running in the same request
prefetch=[
models.Prefetch(
... | [
0.060733262449502945,
-0.024656271561980247,
-0.047001663595438004,
-0.04325169697403908,
0.09038956463336945,
-0.051670223474502563,
0.033108606934547424,
0.010122310370206833,
0.05373125150799751,
-0.030496055260300636,
0.03267873078584671,
0.028486447408795357,
0.02964499592781067,
-0.0... |
d27652f5-a647-5913-838b-dea02233a0ed | /documentation/send-data/databricks/ | Databricks Ingestion | Set Up a Databricks Project | documentation | false | Databricks Ingestion > Set Up a Databricks Project
- Set up [Databricks dedicated compute](https://docs.databricks.com/aws/en/compute/dedicated-overview) with [Databricks Runtime 16.4 LTS, Scala 2.13 variant](https://docs.databricks.com/aws/en/release-notes/runtime/16.4lts), which includes Spark 3.5.2 and Python 3.12.... | [
0.08114506304264069,
-0.027786029502749443,
-0.06753499805927277,
0.010796276852488518,
0.08005299419164658,
-0.03788141906261444,
0.044912002980709076,
0.03569512441754341,
0.014560462906956673,
-0.037026189267635345,
0.005149232689291239,
0.010741555131971836,
0.030126435682177544,
-0.05... |
d2895af3-bad2-5701-ac6a-78c47d387d9e | /documentation/manage-data/collections/ | Collections | Collection Aliases | documentation | true | Collections > Collection Aliases
In a production environment, it is sometimes necessary to switch different versions of vectors seamlessly.
For example, when upgrading to a new version of the neural network.
There is no way to stop the service and rebuild the collection with new vectors in these situations.
Aliases a... | [
0.03736906126141548,
-0.03851202875375748,
-0.049026019871234894,
-0.053225185722112656,
0.017586732283234596,
-0.033257000148296356,
0.034660615026950836,
0.06931278854608536,
0.05911976844072342,
-0.066002257168293,
0.04005253314971924,
0.07023295015096664,
0.05934040620923042,
-0.041714... |
d295b01e-d2a9-5fcd-9e13-41d25de5c699 | /documentation/tutorials-operations/scaling-and-monitoring-pods/ | Scaling and Monitoring Hybrid Cloud | Scale CPU, Memory, and Disk | documentation | true | Scaling and Monitoring Hybrid Cloud > Scale CPU, Memory, and Disk
Start by looking at what your cluster has today:
```bash
CLUSTER_ID="<your-cluster-id>" # from the Cloud console URL
KUBENS="<your-namespace>" # the Kubernetes namespace configured for this Hybrid Cloud environment
qcloud cluster describe $CLUSTER_ID
... | [
0.04940732195973396,
0.0005018556257709861,
-0.045054689049720764,
-0.04293472692370415,
0.06490708142518997,
-0.018259091302752495,
0.058339789509773254,
0.030621234327554703,
0.09521117061376572,
-0.04188079386949539,
0.025102058425545692,
-0.009495511651039124,
0.03089318983256817,
-0.0... |
d2defe97-2a14-54e2-9dfd-49520b384413 | /articles/dedicated-vector-search/ | Why Vector Search Needs a Dedicated Database | Vectors > Vectors are a Transformation | article | false | Why Vector Search Needs a Dedicated Database > Vectors > Vectors are a Transformation
Vectors are obtained from some other source-of-truth data. They can be restored if lost with the same embedding model previously used. At the same time, even small changes in that model can shift the geometry of the vector space, so ... | [
0.013228495605289936,
-0.06934665888547897,
-0.0396716333925724,
-0.03331129252910614,
0.02146976999938488,
-0.008583461865782738,
0.08154363930225372,
0.035844482481479645,
0.07061285525560379,
-0.0187578946352005,
0.007368070539087057,
0.07370366901159286,
0.03604581207036972,
-0.0354451... |
d303b209-7e30-5005-a375-1452ff7b6e48 | /course/essentials/day-5/pitstop-project/ | Project: Building a Recommendation System | What You'll Build | course | false | Project: Building a Recommendation System > What You'll Build
A hybrid recommendation system using:
- **Multi-vector architecture** with dense, sparse, and ColBERT vectors
- **Universal Query API** for atomic multi-stage search
- **RRF fusion** for combining candidates
- **ColBERT reranking** for fine-grained relevan... | [
0.0074950591661036015,
-0.023530468344688416,
-0.07303053885698318,
-0.06340835988521576,
0.07210720330476761,
-0.025315063074231148,
0.07228417694568634,
0.04315204918384552,
0.062003783881664276,
-0.023494994267821312,
0.06474079936742783,
-0.01917365752160549,
0.0977296456694603,
-0.040... |
d30f2a4f-fa8b-5903-b827-b79dcbe2d39b | /documentation/inference/ | Inference | Choose Your Approach | documentation | false | Inference > Choose Your Approach
The right option depends on your deployment and what you need to embed. Use this table to find the best fit for your use case.
| If you… | Use… |
|---|---|
| Need keyword-based retrieval with BM25 | [BM25 on a full-text index](/documentation/search/text-search/full-text-search/#bm25) ... | [
0.02942858450114727,
-0.03679119423031807,
-0.018971741199493408,
-0.05842176824808121,
0.09151913970708847,
-0.029101815074682236,
0.07298453897237778,
0.07009544968605042,
0.0568719245493412,
-0.024049468338489532,
0.03174944594502449,
0.01472531445324421,
0.022925451397895813,
-0.061066... |
d330789c-ea91-5bf2-a985-f89112fec0f1 | /course/essentials/day-1/movie-search-system/ | Demo: Semantic Movie Search | What You've Learned | course | false | Demo: Semantic Movie Search > What You've Learned
This demo brings together every concept from Day 1:
**Chunking in Action:**
You've seen how different chunking strategies affect search quality. Fixed chunking is fast but crude, sentence chunking preserves readability, and semantic chunking captures meaning - but at ... | [
0.06080101057887077,
-0.030003219842910767,
-0.09048061817884445,
-0.03134063631296158,
0.047562453895807266,
-0.04471193626523018,
0.023304009810090065,
0.03402971848845482,
0.07467113435268402,
-0.031382009387016296,
0.05514908209443092,
0.027937473729252815,
0.07536773383617401,
-0.0190... |
d349694e-961c-5a5a-8877-e22660ed2767 | /documentation/tutorials-build-essentials/question-answering-langchain/ | Question Answering with LangChain and Qdrant | documentation | false | Question Answering with LangChain and Qdrant
| Time: 20 min | Level: Beginner | [](https://colab.research.google.com/github/qdrant/examples/blob/master/langchain-integration/langchain_integration.ipynb) |
| --- | ----------- | ----------- |
Bui... | [
-0.014142990112304688,
0.006376405246555805,
-0.039210978895425797,
-0.003752977354452014,
0.07578687369823456,
-0.07412070780992508,
0.06896502524614334,
0.028857732191681862,
0.061069782823324203,
-0.032160282135009766,
0.026450546458363533,
-0.008498179726302624,
0.05845765396952629,
-0... | |
d35acdf3-1756-5271-8946-aa92b7fc6837 | /documentation/tutorials-search-engineering/slicing-filter/ | Deterministic Collection Slicing | Scrolling a Single Slice | documentation | true | Deterministic Collection Slicing > Scrolling a Single Slice
A `SliceCondition` takes an `index` and a `total`. A single `scroll` call only returns up to `limit` points, so to fetch a whole slice you page through it: keep calling `scroll` with the `next_page_offset` from the previous response until it comes back `None`... | [
0.00004655493830796331,
-0.010386887937784195,
-0.056357935070991516,
-0.05358124524354935,
0.0209435373544693,
-0.043138399720191956,
0.025640230625867844,
0.03881852701306343,
0.09510929137468338,
0.0013318429701030254,
0.0013196421787142754,
0.0519813634455204,
0.029612451791763306,
-0.... |
d3667135-e16a-578c-896c-a2f1a93592bf | /articles/agentic-rag/ | What is Agentic RAG? Building Agents with Qdrant | article | false | What is Agentic RAG? Building Agents with Qdrant
Standard [Retrieval Augmented Generation](/articles/what-is-rag-in-ai/) follows a predictable, linear path: receive
a query, retrieve relevant documents, and generate a response. In many cases that might be enough to solve a particular
problem. In the worst case scena... | [
-0.0202427189797163,
-0.05413731187582016,
-0.07133251428604126,
-0.043712299317121506,
0.06597960740327835,
-0.04451635479927063,
0.06047309190034866,
0.033733587712049484,
0.06810739636421204,
-0.03159116953611374,
0.02920214645564556,
0.03896497189998627,
0.04142571613192558,
-0.0236125... | |
d37f5035-4f65-54fc-9314-423d0951d717 | /course/essentials/day-2/collection-tuning-demo/ | Demo: HNSW Performance Tuning | What You’ll Do | course | false | Demo: HNSW Performance Tuning > What You’ll Do
Yesterday you learned the theory behind HNSW indexing. Today you'll see it in action on a 100,000-vector dataset, measuring performance differences and applying optimization strategies that work in production.
**You'll learn to:**
- Optimize bulk upload speed with strate... | [
0.03950237110257149,
-0.021976729854941368,
-0.05930802971124649,
-0.025968661531805992,
0.04244939982891083,
-0.020796198397874832,
0.02154102921485901,
0.02127852849662304,
0.07580186426639557,
-0.012782257981598377,
0.022862937301397324,
0.047120802104473114,
0.044144146144390106,
-0.03... |
d3853376-05d2-5af2-911a-72f7fa2cfde7 | /course/essentials/day-3/sparse-vectors/ | Sparse Vectors and Inverted Indexes | Sparse Vector Representations > Representations as Index-Value Pairs | course | true | Sparse Vectors and Inverted Indexes > Sparse Vector Representations > Representations as Index-Value Pairs
Storing thousands of zeros contributing no information is wasteful.
Instead, a sparse vector can be stored compactly as **(index, value)** pairs of its non-zero entries.
Using the movie example:
```text
[0, 0... | [
0.0671706274151802,
-0.028041766956448555,
-0.03301423043012619,
-0.03795617073774338,
0.04941537603735924,
-0.05789725109934807,
0.040377162396907806,
0.01341932825744152,
0.06356775760650635,
0.0006678432691842318,
-0.016780603677034378,
0.010386994108557701,
0.061201754957437515,
-0.008... |
d3956cc8-8e54-5b76-ab8a-7f251926cdc8 | /documentation/tutorials-basics/generate-embeddings-fastembed/ | FastEmbed: Qdrant's Efficient Python Library for Embedding Generation | Use FastEmbed with Qdrant | documentation | true | FastEmbed: Qdrant's Efficient Python Library for Embedding Generation > Use FastEmbed with Qdrant
Passing the model to `qdrant-client` lets it embed documents and queries for you, so you never handle the vectors yourself.
1. **Create a client.** An in-memory client needs no server:
```python
from qdrant_client i... | [
0.03411959856748581,
-0.04028702154755592,
-0.03120766021311283,
-0.02316644787788391,
0.09025348722934723,
-0.028070857748389244,
0.07317912578582764,
-0.003053434193134308,
0.0574922040104866,
-0.028950579464435577,
0.012175134383141994,
0.008489003404974937,
0.03846825286746025,
-0.0250... |
d397452b-f88d-5def-968d-3d924db5c2de | /documentation/tutorials-basics/reranking-hybrid-search/ | Hybrid Search with Reranking | Implementation > Retrieval | documentation | true | Hybrid Search with Reranking > Implementation > Retrieval
| Position | Title | Description |
|----------|-------|-------------|
| 1 | Station Eleven | A traveling symphony roams a post-pandemic North America. |
| 2 | Slaughterhouse-Five | A nonlinear, time-tripping reflection on war and fate. |
| 3 | The Time Machine ... | [
0.0558909997344017,
-0.013285565190017223,
-0.06917714327573776,
-0.0457230769097805,
0.09061280637979507,
-0.020190933719277382,
0.029158327728509903,
0.01737886480987072,
0.06255800276994705,
0.0055590528063476086,
0.02448403276503086,
-0.03789873793721199,
0.05950704216957092,
-0.015742... |
d3a403fc-eb46-53a8-a71e-ec3fa6c4bcb7 | /articles/what-is-a-vector-database/ | What is a Vector Database? | Hybrid Search | article | false | What is a Vector Database? > Hybrid Search
Sometimes context alone isn’t enough. Sometimes you need precision, too. Dense vectors (embeddings we've talked about so far) are fantastic when you need to retrieve results based on the context or meaning behind the data. Sparse vectors are useful when you also need **keywor... | [
0.004973789677023888,
-0.017660176381468773,
-0.08237199485301971,
-0.04600357636809349,
0.054835204035043716,
-0.027051853016018867,
0.05696886405348778,
0.0029403704684227705,
0.08494408428668976,
-0.016874535009264946,
0.020764658227562904,
0.036560218781232834,
0.06956604868173599,
-0.... |
d3b0def3-fae3-59ff-bd1b-cfe8b5e9cc66 | /documentation/examples/qdrant-dspy-medicalbot/ | Building a Chain-of-Thought Medical Chatbot with Qdrant and DSPy | Dataset Preparation and Indexing | documentation | true | Building a Chain-of-Thought Medical Chatbot with Qdrant and DSPy > Dataset Preparation and Indexing
A medical chatbot is only as good as the knowledge it has access to. For this project, we will leverage the [MIRIAD medical dataset](https://huggingface.co/datasets/miriad/miriad-5.8M), a large-scale collection of medic... | [
0.019365211948752403,
-0.027243953198194504,
-0.05292591080069542,
-0.03255973756313324,
0.0819764956831932,
-0.03433719277381897,
0.07062793523073196,
0.07134480774402618,
0.050819870084524155,
-0.03556857258081436,
0.03399486467242241,
0.04053616151213646,
0.02330736629664898,
-0.0549592... |
d3b751cc-b8cd-58d0-8c7e-e01e3ccbf32a | /articles/metric-learning-tips/ | Metric Learning Tips & Tricks | What is metric learning and why using it? | article | false | Metric Learning Tips & Tricks > What is metric learning and why using it?
According to Wikipedia, metric learning is the task of learning a distance function over objects.
In practice, it means that we can train a model that tells a number for any pair of given objects.
And this number should represent a degree or sco... | [
0.05926045402884483,
-0.006059701554477215,
-0.05131669342517853,
0.010829966515302658,
0.056805044412612915,
0.002556453924626112,
0.0724191814661026,
0.07029307633638382,
0.08709786832332611,
-0.002437941962853074,
0.02497548796236515,
-0.016583876684308052,
0.05833777040243149,
-0.01030... |
d3d70a04-fd85-5665-8fa1-b6385b86e108 | /documentation/search/low-latency-search/ | Low-Latency Search | Use Delayed Fan-Outs | documentation | true | Low-Latency Search > Use Delayed Fan-Outs
*Available as of v1.17.0*
By default, a search operation queries a single replica of each shard in a collection. If a replica responds slowly due to load or network issues, overall search latency can increase. This phenomenon, where a single slow replica increases the 95th or... | [
0.024565910920500755,
-0.061037931591272354,
-0.06457380205392838,
-0.02202398143708706,
0.10070579499006271,
-0.05667196959257126,
0.036585163325071335,
0.03944110870361328,
0.06730427592992783,
-0.030467329546809196,
0.05714233219623566,
-0.0009363624849356711,
0.0500570610165596,
-0.004... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.