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 |
|---|---|---|---|---|---|---|---|
527d99da-c405-56d4-bdfd-ceccf527bcd0 | /documentation/agentic-tools/skills/ | Agent Skills | Installing and Contributing | documentation | false | Agent Skills > Installing and Contributing
For instructions on installing skills in your own coding agent (Cursor, Claude Code, OpenCode, and others), refer to this [README](https://github.com/qdrant/skills/blob/main/README.md).
For guidance on contributing new skills or reporting problems, refer to [CONTRIBUTING.md]... | [
-0.0005456764483824372,
-0.07154912501573563,
-0.060472335666418076,
-0.06393053382635117,
0.05327219143509865,
-0.04532662406563759,
0.05053965374827385,
0.08206788450479507,
0.08032567799091339,
-0.04668831825256348,
0.05706428736448288,
0.0037194695323705673,
0.07054156810045242,
-0.060... |
52ad26a0-5758-5e15-b41c-99e659f1146e | /documentation/manage-data/ | Manage Data | documentation | false | Manage Data
Learn how to structure, store, and organize your data in Qdrant. These pages cover the core building blocks — from individual records and the vectors that represent them, to collections, payloads, and the indexing and quantization options that control how data is stored and retrieved.
## Points
[Points](... | [
0.033306416124105453,
-0.021033475175499916,
-0.050029098987579346,
-0.011940709315240383,
0.04553056135773659,
-0.03630128502845764,
0.08192116767168045,
0.06829699873924255,
0.09473636746406555,
-0.03886215016245842,
0.020539630204439163,
0.053239740431308746,
0.03234853968024254,
-0.016... | |
52c17e9b-b526-582f-af47-8d201a0de92f | /documentation/examples/multitenancy/ | Multitenancy with Qdrant | Setting up the pipeline > Defining a chunking strategy | documentation | true | Multitenancy with Qdrant > Setting up the pipeline > Defining a chunking strategy
Long documents don't fit into a single embedding, so they are usually split into overlapping chunks first.
A few lines of plain Python are enough for a fixed-length splitter. Both values are counted in words, and the
overlap keeps a sent... | [
0.009862350299954414,
-0.052130065858364105,
-0.04365412890911102,
-0.04334811866283417,
0.0607919804751873,
-0.025748837739229202,
0.045861128717660904,
0.04321007803082466,
0.0793754905462265,
-0.0350198931992054,
0.013794164173305035,
-0.008414783515036106,
0.07645603269338608,
-0.01646... |
52c812f1-6619-5ede-8ea7-46dd51b89a3d | /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 |
|-----------------------------------... | [
0.03498239070177078,
-0.01371214259415865,
-0.03422203287482262,
0.01704714633524418,
0.0674980878829956,
-0.06378563493490219,
0.03881599381566048,
0.08273913711309433,
0.047381434589624405,
-0.04891088604927063,
0.029148608446121216,
0.013519578613340855,
0.024394219741225243,
-0.0228762... |
52e2f225-cfb1-5dbb-91da-21f046d74c5b | /documentation/frameworks/swiftide/ | Swiftide | documentation | true | Swiftide
Swiftide is a Rust library for building LLM applications. It supports everything from simple prompt completions to fast, streaming indexing and querying pipelines, and building composable agents that use tools or call other agents.
## High level features
- Simple primitives for common LLM tasks
- Streaming ... | [
0.029543563723564148,
0.005341843701899052,
-0.08543550223112106,
-0.062453072518110275,
0.05596153810620308,
-0.03857940435409546,
0.08435501158237457,
0.02879556640982628,
0.040082912892103195,
-0.03599115461111069,
0.02863970957696438,
-0.0036549910437315702,
-0.0009654896566644311,
-0.... | |
52e91658-baff-5433-b72b-d41578f6e061 | /documentation/search-tuning/before-tuning-a-qdrant-collection/ | What to Check Before Tuning a Qdrant Collection | The Symptom Tells You Where to Start | documentation | false | What to Check Before Tuning a Qdrant Collection > The Symptom Tells You Where to Start
Start with the failure mode, not the config reference. The table maps each symptom to the first useful check and the article that covers it.
| What You See | First Check | Read Next |
|---|---|---|
| You cannot separate a gain from... | [
0.02688765525817871,
-0.06396651268005371,
-0.06558609753847122,
-0.036345820873975754,
0.05727577582001686,
-0.011293605901300907,
0.04956062138080597,
0.0458524227142334,
0.08615773916244507,
-0.032706793397665024,
0.01607854664325714,
0.027605773881077766,
0.03233833238482475,
-0.041727... |
52ebbc53-0dd1-5fb6-ba4b-ec4c2fe7adc1 | /documentation/tutorials-operations/secure-and-expose-hybrid-cloud/ | Securing and Exposing a Hybrid Cloud Cluster | Expose the Cluster Through an Ingress > Verify | documentation | true | Securing and Exposing a Hybrid Cloud Cluster > Expose the Cluster Through an Ingress > Verify
```bash
curl -ik https://qdrant.<your-ingress-external-ip>.sslip.io/collections -H "api-key: $API_KEY"
# HTTP/2 200
```
This request goes through two TLS connections, each with its own certificate. Between Traefik and Qdrant... | [
0.09189601987600327,
-0.06561878323554993,
-0.0454486608505249,
-0.04379119724035263,
0.04798327758908272,
-0.03274068981409073,
0.05669610947370529,
0.04360651224851608,
0.008762006647884846,
-0.02706625685095787,
0.015013804659247398,
0.01598968356847763,
0.04051430523395538,
-0.01457849... |
52f48426-d668-5cf1-a0db-506ada872fd8 | /documentation/embeddings/jina-embeddings/ | Jina AI | Embed Images, Video, Audio, and PDFs | documentation | true | Jina AI > Embed Images, Video, Audio, and PDFs
The `omni` models embed images, video, audio, and PDFs into the same vector space as text. Because `jina-embeddings-v5-omni-small` returns the same text vectors as `jina-embeddings-v5-text-small`, this example adds images to the collection from [Embed Documents](#embed-do... | [
0.031184924766421318,
-0.054666079580783844,
-0.05395716056227684,
-0.06588080525398254,
0.09110322594642639,
0.0021775232162326574,
0.05605205520987511,
0.0649523064494133,
0.061739731580019,
-0.0002763632219284773,
0.03409288823604584,
-0.003084583207964897,
0.035452209413051605,
-0.0362... |
5306e79a-0716-5931-aef7-90718952f049 | /documentation/manage-data/indexing/ | Indexing | Full-Text Index > Tokenizers > Code > Try the Code Tokenizer | documentation | false | Indexing > Full-Text Index > Tokenizers > Code > Try the Code Tokenizer
Enter a query to see which code snippets a text or phrase match finds with the `code` tokenizer. You can also edit the snippets.
 is the framework for solving advanced tasks with language models (LMs) and retrieval models (RMs). It unifies techniques for prompting and fine-tuning LMs — and approaches for reasoning, self-improvement, and augmentation with retrieval and tools.
- Provides c... | [
0.04125870764255524,
-0.04567525535821915,
-0.01762452721595764,
-0.02700912207365036,
0.06303559988737106,
-0.04291762411594391,
0.06518329679965973,
0.006310115568339825,
0.07862318307161331,
-0.037143364548683167,
0.015812905505299568,
0.011190391145646572,
0.07128338515758514,
-0.03003... | |
531513b4-40e3-5d76-b59f-523b87e11ebe | /articles/sparse-vectors/ | Understanding SPLADE and Sparse Vectors | Practical implementation in Python > 5. Retrieve and interpret results | article | true | Understanding SPLADE and Sparse Vectors > Practical implementation in Python > 5. Retrieve and interpret results
After setting up the collection and inserting sparse vectors, the next critical step is retrieving and interpreting the results. This process involves executing a search query and then analyzing the returne... | [
0.06050208583474159,
-0.059636231511831284,
-0.019326914101839066,
-0.01944149285554886,
0.04582573473453522,
-0.030421771109104156,
0.07620875537395477,
0.02950553596019745,
0.04396526515483856,
-0.026626115664839745,
0.011611935682594776,
0.06718609482049942,
0.033608004450798035,
-0.003... |
53238a77-55af-57a5-b103-ca8b6e22495b | /documentation/tutorials-operations/large-scale-search/ | Large-Scale Search | Running search requests | documentation | true | Large-Scale Search > Running search requests
Once all the preparations are done, we can run the search requests and evaluate the results.
You can find the full code of the search process in the [eval.py](https://github.com/qdrant/laion-400m-benchmark/blob/master/eval.py)
This script will run 100 search requests with... | [
0.030520714819431305,
-0.052923690527677536,
-0.006343977991491556,
-0.028957705944776535,
0.06159874424338341,
-0.016019759699702263,
0.05217295140028,
0.029553411528468132,
0.08361515402793884,
-0.04626324027776718,
0.0175063107162714,
0.013240491040050983,
0.07556413859128952,
-0.040213... |
53297344-cadc-5bab-948c-adc11af8ac2a | /documentation/tutorials-basics/search-beginners-local/ | Semantic Search 101 | 3. Define Storage Location | documentation | true | Semantic Search 101 > 3. Define Storage Location
You need to tell Qdrant where to store embeddings. This is a basic demo, so your local computer will use its memory as temporary storage.
```python
client = QdrantClient(":memory:")
``` | [
0.03181874379515648,
-0.04288273677229881,
-0.022209249436855316,
-0.05372020602226257,
0.018110651522874832,
-0.02057143673300743,
0.026023350656032562,
0.06476883590221405,
0.10179083049297333,
-0.008608348667621613,
0.006370999850332737,
-0.0004939686041325331,
0.0645875558257103,
-0.02... |
532e61a2-5bbe-516e-bbed-0ffffe7c8468 | /documentation/tutorials-build-essentials/semantic-cache/ | Semantic Caching for RAG | Setup | documentation | true | Semantic Caching for RAG > Setup
### Prerequisites
You need:
- **A free Qdrant Cloud cluster**. Follow the [Cloud Quickstart](/documentation/cloud-quickstart/) and copy the cluster URL and an API key.
- **An OpenRouter API key**. OpenRouter is chosen for convenience: one key here covers the embeddings inference, the... | [
0.06153133139014244,
-0.05594593659043312,
-0.034265197813510895,
-0.053472600877285004,
0.07329775393009186,
-0.010674228891730309,
0.05797037482261658,
0.04725874215364456,
0.07097890228033066,
-0.045667435973882675,
0.0075698066502809525,
0.007983390241861343,
0.04058538004755974,
-0.04... |
53394634-65b2-5084-8132-f3555a99c365 | /course/essentials/day-2/filterable-hnsw/ | Combining Vector Search and Filtering | Key Takeaways | course | false | Combining Vector Search and Filtering > Key Takeaways
**Filterable HNSW is not a separate indexing mechanism**, it extends the HNSW graph by adding extra edges based on stored payload values to maintain traversal performance under filtering constraints.
**The query planner** automatically selects the optimal strategy... | [
0.011364593170583248,
-0.031310874968767166,
-0.06325013935565948,
-0.03689809888601303,
0.05277758091688156,
-0.02717491239309311,
0.06454349309206009,
0.053927138447761536,
0.06791850179433823,
-0.020564351230859756,
0.05302505940198898,
0.03840892016887665,
0.030441707000136375,
-0.0153... |
533f2706-c0a4-5f19-95c0-70b78e680fdc | /course/essentials/day-3/hybrid-search/ | Hybrid Search and the Universal Query API | course | false | Hybrid Search and the Universal Query API
Day 3
Learn how to combine dense and sparse vector search methods to build powerful hybrid search pipelines that serve diverse user needs. | [
0.022390585392713547,
-0.01546439528465271,
-0.053684066981077194,
-0.04140116274356842,
0.07470192015171051,
-0.04489246383309364,
0.061354439705610275,
0.03740723431110382,
0.07732843607664108,
-0.002804787829518318,
0.023569662123918533,
0.06965680420398712,
0.06718162447214127,
-0.0336... | |
53641f95-9f0f-5049-a397-94bc04fe4698 | /documentation/embeddings/twelvelabs/ | TwelveLabs | documentation | false | TwelveLabs
[TwelveLabs](https://docs.twelvelabs.io/v1.3/docs/guides/create-embeddings) provides Marengo, an embedding model that places video, audio, images, text, and documents in one vector space, so a query in one modality can find content in another. There are two ways to use Marengo with Qdrant:
- Call the Twelv... | [
0.04156866669654846,
-0.03628142178058624,
-0.06905017048120499,
-0.03246849775314331,
0.07977137714624405,
-0.04735000431537628,
0.07786697894334793,
0.05298139154911041,
0.044672440737485886,
-0.0211117435246706,
0.039021071046590805,
0.01763043738901615,
0.058155886828899384,
-0.0157015... | |
5381239c-fcdf-597a-89d1-04304e13bbef | /documentation/cloud-security/ | Cloud Security | Terms of Service and Data Processing Agreement | documentation | false | Cloud Security > Terms of Service and Data Processing Agreement
By using Qdrant Cloud, you agree to our [Terms of Service](https://cloud.qdrant.io/service-agreement) and [Privacy Policy](https://cloud.qdrant.io/privacy-policy). For customers subject to GDPR, we also offer a [Data Processing Agreement (DPA)](https://cl... | [
0.02390606515109539,
0.0037420804146677256,
-0.04990847781300545,
-0.05925552546977997,
0.019520273432135582,
-0.04532025009393692,
0.09855643659830093,
0.04398391395807266,
0.05420843884348869,
-0.02074858546257019,
0.00045987009070813656,
0.019760657101869583,
0.08142410218715668,
-0.030... |
53a03e85-f065-5a35-add9-19ef0ef08aa5 | /documentation/edge/edge-api/shard-lifecycle/ | Shard Lifecycle | Inspect the Path and Configuration | documentation | true | Shard Lifecycle > Inspect the Path and Configuration
*Rust only*
Return the shard's directory and its currently resolved configuration.
```rust
pub fn path(&self) -> &Path
pub fn config(&self) -> parking_lot::RwLockReadGuard<'_, EdgeConfig>
```
`config` returns a read guard rather than a copy, so the configuration ... | [
0.03775327652692795,
-0.036453619599342346,
-0.06696128845214844,
-0.015979046002030373,
0.052187878638505936,
-0.049517672508955,
0.04404059797525406,
0.029373595491051674,
0.010722024366259575,
0.015771884471178055,
-0.0022863091435283422,
0.019910970702767372,
0.028888750821352005,
-0.0... |
53be77e6-b3af-5d28-a146-eb3e3f89f27b | /course/multi-vector-search/module-3/multi-stage-retrieval/ | Multi-Stage Retrieval with Universal Query API | Implementing Multi-Stage Retrieval in Qdrant > Understanding the Query Structure | course | false | Multi-Stage Retrieval with Universal Query API > Implementing Multi-Stage Retrieval in Qdrant > Understanding the Query Structure
The key parts of a multi-stage query:
- **`prefetch`**: Defines the first stage search
- Uses `models.Document` to specify the text and embedding model
- The `using` parameter specifie... | [
0.020760104060173035,
-0.05667877942323685,
-0.016077550128102303,
-0.01200805138796568,
0.06458158046007156,
-0.021250925958156586,
0.0675908699631691,
0.018123185262084007,
0.07364402711391449,
-0.013885557651519775,
0.01082336064428091,
0.00315239024348557,
0.04776494577527046,
-0.06293... |
53bf3f87-4734-59c9-87b7-87694bda0767 | /course/essentials/day-1/distance-metrics/ | Distance Metrics | Quick rule of thumb > Dot Product Similarity | course | true | Distance Metrics > Quick rule of thumb > Dot Product Similarity
- **Common Use Cases:** Recommendation systems, matrix factorization, ranking.
- **Focus:** Both Magnitude and Direction.
Dot product similarity is calculated by multiplying the respective values in the two vectors and then summing those products. Unlike... | [
0.03346597030758858,
-0.045959845185279846,
-0.0823436751961708,
-0.042289163917303085,
0.046261776238679886,
-0.014609439298510551,
0.0700395405292511,
0.026334675028920174,
0.10485797375440598,
-0.03184199705719948,
0.04366448521614075,
-0.0032228431664407253,
0.06762786954641342,
-0.013... |
53cbff53-4d30-5335-a9c2-dedd9411fc55 | /articles/relevance-feedback/ | Relevance Feedback in Qdrant | Experiments > Metric | article | false | Relevance Feedback in Qdrant > Experiments > Metric
We want to check if relevance feedback-based retrieval increases relevance recall. That means surfacing documents more relevant to the query than the vanilla retriever did.
**How did we emulate "we-want-to-surface" quality in experiments?**
* For each query, the re... | [
0.038397885859012604,
-0.03592459857463837,
-0.049415331333875656,
-0.06461729854345322,
0.05024111270904541,
-0.013646828942000866,
0.04140249639749527,
0.011415927670896053,
0.0824275091290474,
-0.0217454731464386,
0.023562977090477943,
0.0065175509080290794,
0.08524630218744278,
0.00489... |
53d09645-0ccb-5d48-9b2e-822337de0f77 | /articles/immutable-data-structures/ | Immutable Data Structures in Qdrant | Conclusion | article | false | Immutable Data Structures in Qdrant > Conclusion
Immutable data structures, while tricky to implement correctly, offer significant performance gains, especially for read-heavy systems like search engines. They allow us to take full advantage of hardware optimizations, reduce memory overhead, and improve cache performa... | [
0.05772462487220764,
-0.041306041181087494,
-0.061938297003507614,
-0.027775729075074196,
0.05992405116558075,
-0.04026981070637703,
0.10865262895822525,
0.03519587963819504,
0.07384593039751053,
0.005648410879075527,
-0.021682627499103546,
0.03116627410054207,
0.044507939368486404,
-0.006... |
541c6d5e-2b66-5a7f-8b2d-bc31ccb5ed9e | /documentation/manage-data/points/ | Points | Conditional Updates | documentation | true | Points > Conditional Updates
_Available as of v1.16.0_
All update operations (including point insertion, vector updates, payload updates, and deletions) support configurable pre-conditions based on filters.
```python
from qdrant_client import QdrantClient, models
client = QdrantClient(url="http://localhost:6333")
... | [
-0.000011662649740173947,
-0.05423429608345032,
-0.04549625143408775,
-0.039688438177108765,
0.03256983309984207,
-0.015528912656009197,
0.032981231808662415,
0.01127590797841549,
0.06148771196603775,
-0.0016670959303155541,
0.026892617344856262,
0.005823620595037937,
0.060453563928604126,
... |
5440bb9f-3c93-5dc9-8e00-64504c534b47 | /course/essentials/day-3/sparse-retrieval-demo/ | Demo: Keyword Search with Sparse Vectors | Sparse Neural Retrieval > SPLADE: Sparse Lexical and Expansion Model > Create a Collection for Sparse Neural Retrieval with SPLADE++ | course | true | Demo: Keyword Search with Sparse Vectors > Sparse Neural Retrieval > SPLADE: Sparse Lexical and Expansion Model > Create a Collection for Sparse Neural Retrieval with SPLADE++
SPLADE models don’t rely on corpus-level statistics like IDF to estimate word relevance. Instead, they generate term weights in sparse represen... | [
0.06924459338188171,
-0.03902989998459816,
-0.015493139624595642,
-0.05842358246445656,
0.05485555902123451,
-0.03537362813949585,
0.060042884200811386,
0.019504308700561523,
0.053544655442237854,
-0.011097156442701817,
0.01764741912484169,
0.014020245522260666,
0.05833526328206062,
-0.003... |
54422dae-c932-51dc-8fc4-631496348f9f | /documentation/frameworks/langgraph/ | LangGraph | documentation | true | LangGraph
[LangGraph](https://github.com/langchain-ai/langgraph) is a library for building stateful, multi-actor applications, ideal for creating agentic workflows. It provides fine-grained control over both the flow and state of your application, crucial for creating reliable agents.
You can define flows that involv... | [
-0.020742520689964294,
-0.024019254371523857,
-0.03020535595715046,
-0.016554921865463257,
0.09587924927473068,
-0.08189798146486282,
0.05318028852343559,
0.0638635978102684,
0.0021734433248639107,
-0.05053436756134033,
0.025700410827994347,
0.020772572606801987,
0.04775639995932579,
-0.05... | |
544afb4c-bca9-527d-98a0-e16782df6ce4 | /documentation/tutorials-basics/cloud-inference-hybrid-search/ | Hybrid Search | Add Data | documentation | true | Hybrid Search > Add Data
Now you can add sample documents, their associated metadata, and a point ID for each. Here's a sample of the [miriad/miriad-4.4M](https://huggingface.co/datasets/miriad/miriad-4.4M) dataset:
| qa_id | paper_id | question | year | venue | specialty | passage_text |
|---|---|---|---|---|---|---... | [
0.05171509087085724,
0.002583650406450033,
-0.03977678716182709,
-0.02513040602207184,
0.07419587671756744,
-0.018930582329630852,
0.05112413316965103,
0.0799235925078392,
0.06170283257961273,
-0.034879375249147415,
0.0429503433406353,
0.055920783430337906,
0.00878281332552433,
-0.01294477... |
544f8492-5dc1-5268-9669-2b0976ce34c0 | /course/essentials/day-3/hybrid-search-demo/ | Demo: Implementing a Hybrid Search System | Next Steps & Resources | course | false | Demo: Implementing a Hybrid Search System > Next Steps & Resources
**What's next:**
1. **Experiment with parameters**: Adjust prefetch limits, try different embedding models
2. **Add reranking**: Explore more complex models for final reranking stage, such as cross-encoders
3. **Build ground truth**: Create evaluation ... | [
0.03646211698651314,
-0.037730176001787186,
-0.041435662657022476,
-0.026875408366322517,
0.07269788533449173,
-0.0481441356241703,
0.03742290288209915,
0.018681518733501434,
0.10181733220815659,
-0.022600237280130386,
0.03202694281935692,
0.009804152883589268,
0.0831465870141983,
-0.05641... |
5466b421-266a-5f3b-b2dc-68ecdb6c9f70 | /documentation/tutorials-build-essentials/video-anomaly-edge-part-3/ | Video Anomaly Detection Part 3: Scoring, Governance, and Deployment | Anomaly Scoring and Incident Formation | documentation | true | Video Anomaly Detection Part 3: Scoring, Governance, and Deployment > Anomaly Scoring and Incident Formation
Raw kNN scores are noisy. A single unusual frame does not constitute an incident. The scoring pipeline applies three layers of processing.
### 1. Temporal Smoothing
An exponential moving average (EMA) dampens... | [
0.04223203286528587,
-0.03916986286640167,
-0.062140755355358124,
-0.05914654582738876,
0.030626025050878525,
0.011876667849719524,
0.027616648003458977,
0.03586706519126892,
0.08100869506597519,
-0.022600822150707245,
0.01968853920698166,
0.016771331429481506,
0.07695615291595459,
-0.0192... |
547a97cf-d6fd-5312-87a0-daf008ad60c0 | /documentation/manage-data/collections/ | Collections | Create a Collection > Vector Datatypes | documentation | true | Collections > Create a Collection > Vector Datatypes
*Available as of v1.9.0*
By default, Qdrant stores each vector dimension as a 32-bit float. Memory and storage grow linearly with dimensionality, so for large vectors this adds up quickly. To reduce that cost, or to store vectors that are already lower precision, y... | [
0.03969945013523102,
-0.0433209128677845,
-0.03233158960938454,
-0.028414525091648102,
0.02680487558245659,
-0.05755740776658058,
0.037679560482501984,
0.045339953154325485,
0.09841493517160416,
-0.03507527709007263,
0.054129429161548615,
0.036029402166604996,
0.05095084756612778,
-0.01492... |
54a09b8a-b9b0-5d6b-a3d5-9338823ba2cb | /documentation/manage-data/points/ | Points | Awaiting Result | documentation | true | Points > Awaiting Result
If the API is called with the `&wait=false` parameter, or if it is not explicitly specified, the client will receive an acknowledgment of receiving data:
```json
{
"result": {
"operation_id": 123,
"status": "acknowledged"
},
"status": "ok",
"time": 0.000206061
}
```
This resp... | [
-0.022621409967541695,
-0.07139071077108383,
0.019546855241060257,
-0.057217203080654144,
0.05121077969670296,
-0.04651789367198944,
0.07005589455366135,
0.009664982557296753,
0.04096393287181854,
-0.009467599913477898,
-0.0014811070868745446,
0.0840972512960434,
0.04114893078804016,
-0.01... |
54a76f83-696d-56e7-886c-2c535a090663 | /documentation/cloud-rbac/role-management/ | Role Management | documentation | false | Role Management
> 💡 You can access this in **Access Management > User & Role Management** *see [this page for details](/documentation/cloud-rbac/).*
A **Role** contains a set of **permissions** that define the ability to perform or control specific actions in Qdrant Cloud. Permissions are accessible through the Perm... | [
0.009852519258856773,
-0.010755591094493866,
-0.044300567358732224,
-0.06368528306484222,
0.052173320204019547,
-0.03239966556429863,
0.05825857073068619,
0.08431641012430191,
0.1037454828619957,
-0.03976675868034363,
0.019933586940169334,
-0.017273155972361565,
0.0848320946097374,
-0.0361... | |
54ad1d20-8ee8-5342-a807-df96eddb211b | /documentation/platforms/pipedream/ | Pipedream | documentation | false | Pipedream
[Pipedream](https://pipedream.com/) is a development platform that allows developers to connect many different applications, data sources, and APIs in order to build automated cross-platform workflows. It also offers code-level control with Node.js, Python, Go, or Bash if required.
You can use the [Qdrant a... | [
0.009825551882386208,
-0.06256214529275894,
-0.02952725812792778,
-0.03679371625185013,
0.04717317223548889,
-0.04250558465719223,
0.0859927162528038,
0.043392110615968704,
0.036930713802576065,
-0.0254118200391531,
0.008605421520769596,
0.01880110800266266,
0.017988121137022972,
-0.047802... | |
54c54d06-1209-51dd-9363-973c72ff40f2 | /documentation/scaling/horizontal-scaling/ | Horizontal Scaling | Where to Go Next | documentation | false | Horizontal Scaling > Where to Go Next
- [Resilience](/documentation/scaling/resilience/) covers what these mechanics buy you in fault tolerance, Multi-AZ, and failover.
- [Distributed Deployment](/documentation/scaling/distributed_deployment/) covers the practical configuration: enabling distributed mode, sharding, re... | [
0.05536866560578346,
-0.0003242476377636194,
-0.05702269449830055,
-0.04301882162690163,
0.09254666417837143,
-0.025487083941698074,
0.06454961001873016,
0.03315672278404236,
0.04902224242687225,
-0.0032118232920765877,
0.06674355268478394,
-0.007737588603049517,
0.03249755874276161,
-0.01... |
54c81b6e-7535-5246-92a3-9d2666c6093a | /articles/minicoil-en-es/ | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier | From Words to Concepts | article | false | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier > From Words to Concepts
That's how a concept of a **concept** was born (*applause*): a small set of English and Spanish words that translate to each other, such as `{cat, gato}` or `{bat, murciélago, bate}`.
> Concepts are not specific to English an... | [
0.02861669287085533,
-0.03508066385984421,
-0.05089915543794632,
-0.05275058373808861,
0.09522845596075058,
-0.06292528659105301,
0.03218149021267891,
0.04606514796614647,
0.05525229871273041,
-0.026310984045267105,
0.01571858488023281,
0.02311095781624317,
0.044492464512586594,
-0.0064341... |
54d9faf9-0af1-50c7-bda5-751eea74cf6e | /course/essentials/day-4/what-is-quantization/ | Vector Quantization Methods | Binary Quantization | course | true | Vector Quantization Methods > Binary Quantization
[Binary quantization](https://youtu.be/wRnSDDzLQmk) represents the extreme compression approach, reducing each dimension to a single bit through sign-based thresholding: values greater than zero become 1, values less than or equal to zero become 0. This transforms a 15... | [
0.06018097326159477,
-0.05551749840378761,
-0.05588271841406822,
-0.050766509026288986,
0.055767208337783813,
-0.03017544001340866,
0.0463540181517601,
0.034302134066820145,
0.1218322366476059,
-0.032406311482191086,
0.040239229798316956,
-0.0011684155324473977,
0.038780372589826584,
-0.02... |
54e03122-42b6-55cf-b558-ff0e8bce4ad9 | /documentation/tutorials-develop/code-search/ | Semantic Search for Code | Querying the codebase > Grouping the results | documentation | true | Semantic Search for Code > Querying the codebase > Grouping the results
You can improve the search results, by grouping them by payload properties.
In our case, we can group the results by the module. If we use code embeddings,
we can see multiple results from the `map_index` module. Let's group the
results and assume... | [
0.05312034860253334,
-0.04463335871696472,
-0.03886190056800842,
-0.0026959904935210943,
0.0652129203081131,
-0.018353931605815887,
0.06351372599601746,
0.06255177408456802,
0.05789649114012718,
-0.013002783060073853,
0.05478259176015854,
0.0219977218657732,
0.02967749536037445,
-0.0414506... |
54e52609-ff0a-5b49-aae7-248832cb31ba | /course/multi-vector-search/module-2/how-colpali-works/ | How ColPali Models Work | The Vision Language Model Foundation > The Two-Component Architecture | course | false | How ColPali Models Work > The Vision Language Model Foundation > The Two-Component Architecture
Think of ColPali's architecture as having "eyes" and a "brain":
**Vision Encoder (SigLIP) - The Eyes:**
- Takes the document image and divides it into patches (the 32×32 grid we'll explore next)
- Processes each patch thro... | [
0.016069218516349792,
-0.04056958481669426,
-0.049733731895685196,
-0.04071762040257454,
0.07581683248281479,
0.0011378660565242171,
0.04047833010554314,
0.036349814385175705,
0.03015788085758686,
-0.03186342120170593,
0.03162335604429245,
-0.008131938986480236,
0.0716596469283104,
-0.0200... |
54fc26bb-102a-525e-9e80-73a3a6620bbb | /documentation/migrate-to-qdrant/from-weaviate/ | From Weaviate | Pre-Create Your Qdrant Collection | documentation | true | From Weaviate > Pre-Create Your Qdrant Collection
```bash
curl -X PUT 'https://your-instance.cloud.qdrant.io:6333/collections/your-collection' \
-H 'api-key: your-qdrant-api-key' \
-H 'Content-Type: application/json' \
-d '{
"vectors": {
"size": 384,
"distance": "Cosine"
... | [
0.034566882997751236,
-0.04232967644929886,
-0.031764835119247437,
-0.01834939606487751,
0.06610789895057678,
-0.02261463925242424,
0.02617572993040085,
0.06497185677289963,
0.059858452528715134,
-0.03925471752882004,
0.015366572886705399,
0.017872624099254608,
0.04629572108387947,
0.01633... |
550aec51-fa64-55df-9e47-770f6b8e8452 | /articles/minicoil-en-es/ | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier | article | false | miniCOIL EN-ES: Sparse Neural Retrieval Across the Language Barrier
*Sparse neural retrieval finally started getting more and more attention ([SPARSEUP](https://www.linkup.so/blog/introducing-sparseup-by-linkup), [MILCO](https://arxiv.org/abs/2510.00671), [sparse encoders in Sentence Transformers](https://huggingface.... | [
0.05202040076255798,
-0.046811360865831375,
-0.035772114992141724,
-0.045784126967191696,
0.056733615696430206,
-0.05294553190469742,
0.06906449794769287,
0.02087508887052536,
0.05698682740330696,
-0.035586412996053696,
0.014962708577513695,
0.04306037724018097,
0.06938587129116058,
-0.013... | |
550e8d52-bc01-5072-b6be-1ed5dd7d660c | /documentation/migration-guidance/data-integrity/ | Data Integrity | documentation | false | Data Integrity
Once you've established a [baseline](/documentation/migration-guidance/pre-migration-baseline/), you first need to check data integrity. Data integrity answers the question: "Did all my data arrive, and did it arrive correctly?" These are the fastest checks to run and catch the most common migration fai... | [
0.025941291823983192,
0.007098147179931402,
-0.034957531839609146,
-0.06610741466283798,
0.00925495382398367,
-0.043677784502506256,
0.04888791963458061,
0.02556149661540985,
0.03590911626815796,
0.0004674431693274528,
0.022656673565506935,
0.058964040130376816,
0.044433776289224625,
-0.03... | |
554e7fba-1fdf-59ca-bcd0-e94d68920a9b | /documentation/search/text-search/ | Text Search | Lexical Search | documentation | true | Text Search > Lexical Search
Lexical search, also known as keyword-based search, is a traditional search technique that relies on matching words or phrases in the text. Many applications require a combination of semantic and traditional lexical search. A good example is in e-commerce, where users may want to search fo... | [
0.029126983135938644,
-0.04619121551513672,
-0.0683702602982521,
-0.02798646129667759,
0.06781075894832611,
-0.03923376649618149,
0.05963446944952011,
-0.01766720600426197,
0.06625343859195709,
-0.004429622087627649,
0.04949570819735527,
0.023837007582187653,
0.054426390677690506,
-0.00614... |
55513402-86c3-5bf6-b7ce-d55322324fc5 | /documentation/hybrid-cloud/networking-logging-monitoring/ | Networking, Logging & Monitoring | Configure network policies | documentation | true | Networking, Logging & Monitoring > Configure network policies
For security reasons, each database cluster is secured with network policies. By default, database pods only allow egress traffic between each other and ingress traffic to ports 6333 (REST) and 6334 (gRPC) from within the Kubernetes cluster.
You can modify... | [
0.03922363370656967,
-0.006640139501541853,
-0.0548405759036541,
-0.010058757849037647,
0.036530014127492905,
-0.06698527932167053,
0.08713123202323914,
0.07899153232574463,
0.04240201413631439,
0.00918256863951683,
0.03371860831975937,
0.012047307565808296,
0.019091622903943062,
-0.023813... |
556bacf0-6344-56f0-8e17-4e13760dc17c | /documentation/tutorials-build-essentials/agentic-rag-crewai-zoom/ | Agentic RAG with CrewAI | Usage > Building the AI Agent System | documentation | true | Agentic RAG with CrewAI > Usage > Building the AI Agent System
Our AI system uses a tool-based approach. Let's start with the simplest tool - a calculator for meeting statistics:
```python
class CalculatorTool(BaseTool):
name: str = "calculator"
description: str = "Perform basic mathematical calculations"
... | [
0.014051135629415512,
-0.04941355064511299,
-0.05897267535328865,
-0.030620718374848366,
0.07251373678445816,
-0.03959810361266136,
0.06396942585706711,
0.03642246499657631,
0.06647781282663345,
-0.029070015996694565,
0.03816351294517517,
-0.009511876851320267,
0.0526171438395977,
-0.03627... |
558807a5-2c86-5d09-9ba6-c23bf30ba899 | /course/essentials/day-6/final-project/ | Final Project: Production-Ready Documentation Search Engine | Build Steps > Step 5: Search Pipeline | course | false | Final Project: Production-Ready Documentation Search Engine > Build Steps > Step 5: Search Pipeline
Convert user queries into the three vector representations needed for hybrid search. Use Qdrant’s Universal Query API to combine dense and sparse with server-side fusion.
**Multistage pipeline**
* **Stage 1 – Hybrid r... | [
0.03501247242093086,
-0.042370423674583435,
-0.04576515033841133,
-0.05643955618143082,
0.07262378185987473,
-0.050264012068510056,
0.0807153731584549,
0.03906790539622307,
0.05696553364396095,
-0.03660053759813309,
0.03077549859881401,
0.019223151728510857,
0.05818738043308258,
-0.0219163... |
558ae4b7-56f8-5943-85b2-e7979dc1f566 | /documentation/cloud/cluster-monitoring/ | Monitor Clusters | Qdrant Database Metrics and Telemetry | documentation | false | Monitor Clusters > Qdrant Database Metrics and Telemetry
You can also directly access the metrics and telemetry that the Qdrant database nodes provide.
To scrape metrics from a Qdrant cluster running in Qdrant Cloud, an [API key](/documentation/cloud/authentication/) is required to access `/metrics` and `/sys_metrics... | [
0.013393367640674114,
-0.017933668568730354,
-0.031418684870004654,
-0.01116044819355011,
0.047331802546978,
-0.06188970059156418,
0.0918828472495079,
0.07560510188341141,
0.07088828831911087,
-0.04656730219721794,
0.051874853670597076,
-0.0028591004665941,
0.06863494962453842,
-0.02343933... |
55c1c0d8-8a3d-517c-be97-f39801e8a169 | /documentation/private-cloud/api-reference/ | API Reference | qdrant.io/v1 > Resource Types > AuditRotation | documentation | false | API Reference > qdrant.io/v1 > Resource Types > AuditRotation
_Underlying type:_ _string_
AuditRotation specifies the rotation interval for audit log files.
_Validation:_
- Enum: [daily hourly]
_Appears in:_
- [AuditConfig](#auditconfig)
| Field | Description |
| --- | --- |
| `daily` | |
| `hourly` | | | [
-0.008187178522348404,
-0.05759434774518013,
-0.014129191637039185,
-0.0778517946600914,
0.0668843612074852,
-0.04109383374452591,
-0.003460078500211239,
0.09902513772249222,
0.09662438929080963,
-0.025703540071845055,
0.029995566233992577,
-0.002055950928479433,
0.036486685276031494,
-0.0... |
55c29e74-aa6b-5eab-8527-38b928615e0c | /documentation/fastembed/fastembed-minicoil/ | Working with miniCOIL | Convert to Sparse Vectors & Upload to Qdrant | documentation | true | Working with miniCOIL > Convert to Sparse Vectors & Upload to Qdrant
Next, we need to convert titles to miniCOIL sparse representations and upsert them into the configured collection.
Qdrant and FastEmbed integration allows for hiding the inference process under the hood.
That means:
- FastEmbed downloads the selec... | [
0.06497740745544434,
-0.07183675467967987,
-0.019831271842122078,
-0.04659033939242363,
0.061948440968990326,
-0.04112202301621437,
0.11881785839796066,
0.03879706934094429,
0.02980867587029934,
-0.008959674276411533,
0.021180182695388794,
0.03259456157684326,
0.03449583798646927,
-0.03568... |
55c6da01-1682-5f6d-81fe-a21f408ce276 | /documentation/private-cloud/api-reference/ | API Reference | qdrant.io/v1 > Resource Types > NodeInfo | documentation | false | API Reference > qdrant.io/v1 > Resource Types > NodeInfo
_Appears in:_
- [QdrantCloudRegionStatus](#qdrantcloudregionstatus)
| Field | Description | Default | Validation |
| --- | --- | --- | --- |
| `name` _string_ | Name specifies the name of the node | | |
| `region` _string_ | Region specifies the region of the... | [
0.0306719858199358,
-0.032774925231933594,
-0.020385269075632095,
-0.05964753404259682,
0.06791090220212936,
-0.049838997423648834,
0.05189324915409088,
0.07001813501119614,
0.05912983417510986,
-0.04391564801335335,
0.041262850165367126,
0.012100985273718834,
0.030536409467458725,
-0.0218... |
55c7614d-0eae-5790-8e9a-b223da73fc0e | /documentation/embeddings/gemini/ | Gemini | Embed Documents | documentation | true | Gemini > Embed Documents
`gemini-embedding-2` combines all the parts of one `Content` object into a single embedding. A plain list of strings is treated as one `Content`, so it returns one aggregated vector. To get one vector per document, wrap each document in its own `Content`. The response lists the embeddings in i... | [
0.020611830055713654,
-0.03640877082943916,
-0.03915994614362717,
-0.051950909197330475,
0.049266181886196136,
-0.005971723701804876,
0.05975315347313881,
0.041323222219944,
0.06853872537612915,
-0.05024610459804535,
0.011447212658822536,
0.04079310968518257,
0.018602052703499794,
-0.02423... |
5619fce8-3c82-5b17-a064-3fc501f06796 | /documentation/tutorials-develop/serverless-semantic-search/ | Serverless Semantic Search | The Lambda Function | documentation | true | Serverless Semantic Search > The Lambda Function
Now join the parts into the Lambda function. The function creates its clients once per execution environment, so later invocations reuse them. It reads its configuration from environment variables, which you set when you deploy. Put the following into `src/bin/search.rs... | [
0.0360301211476326,
-0.05950099602341652,
-0.042869776487350464,
-0.05191249027848244,
0.0752125158905983,
-0.04190808907151222,
0.04009217023849487,
0.04393695667386055,
0.07281851768493652,
-0.04258707910776138,
0.018648767843842506,
0.020813386887311935,
0.014281338080763817,
-0.0089523... |
5633ff03-764e-55d8-a273-54b9a8ba0fd2 | /articles/modern-sparse-neural-retrieval/ | Modern Sparse Neural Retrieval: From DeepCT to SPLADE++ | Sparse Neural Retrieval Evolution > The Pioneer Of Sparse Neural Retrieval | article | false | Modern Sparse Neural Retrieval: From DeepCT to SPLADE++ > Sparse Neural Retrieval Evolution > The Pioneer Of Sparse Neural Retrieval
The authors of one of the first sparse retrievers, the [`Deep Contextualized Term Weighting framework (DeepCT)`](https://arxiv.org/pdf/1910.10687),
predict an integer word’s impact valu... | [
0.03715866059064865,
-0.05135820806026459,
-0.016337255015969276,
-0.03409504517912865,
0.04272305220365524,
-0.01625809259712696,
0.03532164543867111,
0.035385385155677795,
0.0790693610906601,
-0.028725534677505493,
-0.024599120020866394,
0.02804861031472683,
0.046400438994169235,
-0.0497... |
563c4844-35c3-5fc4-a48c-118ee763a6d7 | /course/beginners/module-1/when-a-system-needs-both/ | When a System Needs Both | course | false | When a System Needs Both
Module 1
Those failures show why no single retrieval method works for every query. A search system may need semantic search to find related meaning, filters to match exact values such as a SKU, and keyword search when matching terms should affect the ranking.
**Hybrid search** specifically c... | [
0.007272962015122175,
-0.03708326816558838,
-0.0738828182220459,
-0.037898167967796326,
0.09291253983974457,
-0.014933577738702297,
0.06435732543468475,
-0.0038343050982803106,
0.09693192690610886,
-0.0002204198099207133,
0.02418452501296997,
0.013245618902146816,
0.07835889607667923,
-0.0... | |
5649fb22-5d69-5972-b3c9-da865069ed9c | /documentation/send-data/qdrant-airflow-astronomer/ | Querying with Airflow | Implementation > Add the data corpus | documentation | true | Querying with Airflow > Implementation > Add the data corpus
Let's add some sample data to work with. Paste the following content into a file called `books.txt` file within the `include` directory.
```text
1 | To Kill a Mockingbird (1960) | fiction | Harper Lee's Pulitzer Prize-winning novel explores racial injustice... | [
-0.0028360048308968544,
-0.0038984508719295263,
-0.028096262365579605,
-0.06374633312225342,
0.09187683463096619,
-0.014115151949226856,
-0.03734711557626724,
0.057939838618040085,
0.015077514573931694,
0.003776953089982271,
0.027733679860830307,
0.020937204360961914,
0.05147505924105644,
... |
566adab0-60ed-5bbe-b7c6-090ad5cad09d | /documentation/production-operations/vector-search-in-production/ | Vector Search in Production: Scaling, HA & Tuning Guide | 3. What's the Best Way to Scale the Database and Optimize Resources? > Manage Your Costs by Scaling Up or Down | documentation | false | Vector Search in Production: Scaling, HA & Tuning Guide > 3. What's the Best Way to Scale the Database and Optimize Resources? > Manage Your Costs by Scaling Up or Down
Some teams scale up for daytime surges, then scale down overnight to save resources. If you do this, ensure data is sharded and replicated appropriate... | [
0.03649941459298134,
-0.04248163476586342,
-0.07011697441339493,
-0.0313492976129055,
0.09518714994192123,
-0.04369472712278366,
0.02365720644593239,
0.01606808230280876,
0.06949351727962494,
-0.02796993963420391,
0.02829953283071518,
-0.002567743184044957,
0.019077442586421967,
-0.0418908... |
567348e0-3f91-501b-8fe9-4321da15b65e | /documentation/migration-guidance/data-integrity/ | Data Integrity | 2. Vector Dimension Verification | documentation | true | Data Integrity > 2. Vector Dimension Verification
Confirm that vector dimensions match your source configuration:
```py
collection_info = client.get_collection("your_collection")
qdrant_dim = collection_info.config.params.vectors.size
# For named vectors:
# qdrant_dim = collection_info.config.params.vectors["dense"].... | [
0.0577092319726944,
-0.0389653816819191,
-0.0394306518137455,
-0.06723897159099579,
0.018948357552289963,
-0.004301687702536583,
0.09654387831687927,
0.07374823093414307,
0.05583352595567703,
-0.008792509324848652,
0.009809808805584908,
0.01578819938004017,
0.06193144619464874,
-0.02460065... |
56793824-b321-5655-871d-642474a39425 | /documentation/frameworks/cognee/ | Cognee | documentation | false | Cognee
Embeddings make it easy to retrieve similar chunks of information — but most agent tasks require more: structure, temporal context, and cross-document reasoning. That's where Cognee comes in: it turns raw data sources into AI memory —a semantic data layer based on a modular, queryable knowledge graph backed by ... | [
0.03590532764792442,
-0.0160419549793005,
-0.04426960274577141,
-0.09345048666000366,
0.05433885380625725,
-0.029320217669010162,
0.08509136736392975,
0.06977182626724243,
0.022799314931035042,
0.004700870253145695,
0.04828923940658569,
0.027788827195763588,
0.08021946996450424,
-0.0074998... | |
568a3097-677c-5e20-bc7b-0683adb69317 | /documentation/embeddings/openai/ | OpenAI | Embed Queries | documentation | true | OpenAI > Embed Queries
OpenAI models have no separate query mode. Embed the query with the same model you used for the documents:
```python
query_vector = (
openai_client.embeddings.create(model=MODEL, input="What is Qdrant?")
.data[0]
.embedding
)
hits = client.query_points(
collection_name="{collec... | [
0.027274170890450478,
-0.047697123140096664,
-0.005584025755524635,
-0.04485517740249634,
0.0935898944735527,
0.006152816582471132,
0.022017015144228935,
0.06177537888288498,
0.020508915185928345,
-0.022752735763788223,
0.038729988038539886,
-0.011697730049490929,
0.062261782586574554,
-0.... |
56a26828-a916-56cb-9f44-97260984821a | /articles/hyperbolic-embeddings-qdrant/ | Why Hyperbolic Embeddings Matter for Hierarchical Data | Serving Them with Qdrant | article | true | Why Hyperbolic Embeddings Matter for Hierarchical Data > Serving Them with Qdrant
Getting a good embedding was only half the problem. The next question was how to search it.
Hyperbolic distance can be converted into an inner product by adding two extra dimensions. Under brute force search in Faiss, that worked well. ... | [
0.06928624212741852,
-0.022384265437722206,
-0.02195093221962452,
-0.031064460054039955,
0.04572780057787895,
-0.04504719376564026,
0.0688575729727745,
0.03637846186757088,
0.09447703510522842,
-0.028180642053484917,
0.03323669359087944,
0.016206765547394753,
0.06412260979413986,
-0.020115... |
56a2cb22-d537-5e3f-a2d9-d0494241c31f | /articles/universal-io/ | Universal I/O in Qdrant | article | false | Universal I/O in Qdrant
Most software is built on an unspoken assumption: data lives on disk, and is read from and written to it. The assumption usually holds, until a new way (or need) of storing data appears.
For Qdrant, that happened with [serverless mode](https://qdrant.tech/serverless/), where the data sits in o... | [
0.055590469390153885,
-0.03269175812602043,
-0.07155349850654602,
-0.008755119517445564,
0.07461699098348618,
-0.02280859462916851,
0.04992670565843582,
0.013244809582829475,
0.052308548241853714,
-0.002701655961573124,
0.0190927442163229,
-0.010259822942316532,
0.06956068426370621,
-0.019... | |
56a42869-e4c7-5932-b6d2-b0ce7b161db0 | /documentation/snapshots/ | Snapshots | Storage > Share a bucket between deployments | documentation | true | Snapshots > Storage > Share a bucket between deployments
*Available as of v1.20.0*
Each of `s3_config`, `gcs_config`, and `azure_config` accepts an optional `prefix`. Qdrant stores objects as `<prefix>/snapshots/...`, so several deployments can share one bucket or container without overwriting each other's snapshots.... | [
0.025865891948342323,
-0.053334951400756836,
-0.036818038672208786,
-0.024353159591555595,
0.042564406991004944,
-0.01726303994655609,
0.06565867364406586,
0.027271252125501633,
0.042268816381692886,
-0.028276806697249413,
-0.0005056741647422314,
0.001729245064780116,
0.0307476669549942,
-... |
56a4ff49-6c98-5032-accb-a5c4bfc992af | /documentation/tutorials-operations/blue-green-deployment/ | Blue-Green Cluster Deployment | Rollback | documentation | false | Blue-Green Cluster Deployment > Rollback
If you need to roll back after the cutover:
- If the blue cluster is still running, switch your application config back to the blue cluster's endpoint and API key.
- If you have already decommissioned the blue cluster, restore from a snapshot into a new cluster and point you... | [
0.012290988117456436,
-0.03108801692724228,
0.020714839920401573,
-0.09137070178985596,
0.02834611013531685,
-0.04321625083684921,
0.022937562316656113,
0.01623053289949894,
0.05435694381594658,
-0.018838290125131607,
0.01773875020444393,
0.04311104863882065,
0.01065224688500166,
-0.003987... |
56b91a1d-d978-5190-a14c-13f64352e574 | /documentation/cloud-user-profile/ | User Profile | Security | documentation | false | User Profile > Security
Use the **Security** page to protect your login with multi-factor authentication (MFA). Like the rest of your profile, MFA is a personal setting. It applies to your user wherever you sign in, across every account you belong to.
MFA is optional. Until you set up a method, your login works the s... | [
0.026006482541561127,
-0.04245831072330475,
-0.021667299792170525,
-0.050343237817287445,
0.05620658025145531,
-0.008279683068394661,
-0.007252576760947704,
0.020304085686802864,
0.060284119099378586,
-0.019882386550307274,
0.012009130790829659,
0.037543896585702896,
0.08992944657802582,
-... |
56b98e83-ebd4-5cc7-93cc-55def389acb2 | /articles/dedicated-vector-search/ | Why Vector Search Needs a Dedicated Database | Objections to a Dedicated Service > "I would pay for two systems" | article | false | Why Vector Search Needs a Dedicated Database > Objections to a Dedicated Service > "I would pay for two systems"
You pay for the resources you use, not for the number of systems you run. A dedicated engine lets you tune those resources to the vector workload. [Quantization](/documentation/manage-data/quantization/), f... | [
-0.010677702724933624,
-0.05045240744948387,
-0.05021607130765915,
-0.009825526736676693,
0.04770752787590027,
0.006852415855973959,
0.0514705628156662,
0.04001631215214729,
0.07002536952495575,
0.0051054758951067924,
0.04156966507434845,
0.008522802963852882,
0.03271307423710823,
-0.00529... |
56c5e833-0265-5edb-94bd-8b54caf79991 | /documentation/production-operations/predictable-search-latency/ | Configure Qdrant's Optimizer for Predictable Search Latency | Deferred Indexing Means Optimizing All at Once | documentation | false | Configure Qdrant's Optimizer for Predictable Search Latency > Deferred Indexing Means Optimizing All at Once
As noted earlier, that recovery only applies if a temporary loss of recall for the most recent writes is acceptable in exchange for better query latency. | [
0.011354289017617702,
-0.03316904231905937,
-0.02553475834429264,
-0.04541179537773132,
0.044835012406110764,
-0.04868323355913162,
0.027296781539916992,
0.01205563172698021,
0.09782449901103973,
-0.01441397424787283,
0.04026174172759056,
0.027198629453778267,
0.046338848769664764,
-0.0301... |
56c8f515-00b3-5df6-be15-cab6f593f02d | /documentation/migrate-to-qdrant/from-dynamodb/ | From Amazon DynamoDB | documentation | true | From Amazon DynamoDB
## What You Need from AWS
- **Table name**: the DynamoDB table containing your items
- **Vector index name**: the vector index to migrate
- **AWS credentials**: configured via `aws configure` or environment variables (`AWS_ACCESS_KEY_ID`, `AWS_SECRET_ACCESS_KEY`, `AWS_REGION`)
Set your AWS crede... | [
0.05706816166639328,
-0.03974587097764015,
-0.019109200686216354,
-0.02677418850362301,
0.042646851390600204,
-0.03777112811803818,
0.06282734870910645,
0.03821960836648941,
0.03347622975707054,
-0.03726246953010559,
-0.002233260776847601,
0.032043032348155975,
0.058961402624845505,
-0.028... | |
56e73e89-4526-571d-9a49-9187bb37eddc | /documentation/snapshots/ | Snapshots | Storage > Google Cloud Storage | documentation | true | Snapshots > Storage > Google Cloud Storage
*Available as of v1.20.0*
To store snapshots in Google Cloud Storage (GCS), set `snapshots_storage` to `gcs` and configure `gcs_config` in the [configuration](/documentation/ops-configuration/configuration/) file:
```yaml
storage:
snapshots_config:
# Use 'gcs' to stor... | [
-0.015393786132335663,
-0.005607582628726959,
-0.02211865410208702,
-0.0543014220893383,
0.03151052072644234,
-0.03266696259379387,
0.040124718099832535,
0.024293916299939156,
0.052738696336746216,
-0.05884457752108574,
-0.0026614079251885414,
-0.0017952736234292388,
0.05415138229727745,
-... |
57029cf6-03c6-5d12-a9cb-0d1a83bd8003 | /documentation/tutorials-search-engineering/sparse-embeddings-ecommerce-part-3/ | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 3: Evaluation and Hard Negatives | Indexing Products in Qdrant | documentation | true | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 3: Evaluation and Hard Negatives > Indexing Products in Qdrant
You need Qdrant 1.19 or later (Qdrant Cloud, or `docker run -p 6333:6333 qdrant/qdrant`) and `pip install "qdrant-client>=1.19" sentence-transformers`. Load the published model or your Part 2 check... | [
0.036714110523462296,
-0.04367974400520325,
-0.04540400207042694,
-0.03394434601068497,
0.059610169380903244,
-0.051267631351947784,
0.07500336319208145,
-0.0003871819353662431,
0.07735482603311539,
-0.02179638110101223,
0.00893764104694128,
0.021413633599877357,
0.05390480160713196,
-0.05... |
5703fd5f-8d13-5c34-b8e4-240f03e1b90e | /articles/what-is-vector-quantization/ | What is Vector Quantization? | 3. What is Product Quantization? | article | true | What is Vector Quantization? > 3. What is Product Quantization?

[Product Quantization](https://qdrant.tech/documentation/manage-data/quantization/#product-quantization) is a method used to compress high-dimensional vectors by representing them wi... | [
0.04400760307908058,
-0.03198936581611633,
-0.06782980263233185,
-0.05468422546982765,
0.06339997798204422,
0.01535895373672247,
0.060495659708976746,
0.022435490041971207,
0.12097504734992981,
-0.023900100961327553,
0.026546869426965714,
0.012749492190778255,
0.06302246451377869,
-0.04149... |
5718fdf2-101c-5dc6-b465-790dc93ca1ce | /documentation/tutorials-search-engineering/using-multivector-representations/ | Multivectors and Late Interaction | documentation | false | Multivectors and Late Interaction
| Time: 30 min | Level: Intermediate |
| --- | ----------- |
Multivector Representations are one of the most powerful features of Qdrant. However, most people don't use them effectively, resulting in massive RAM overhead, slow inserts, and wasted compute.
In this tutorial, you'll di... | [
0.012555552646517754,
-0.017454883083701134,
-0.031090719625353813,
-0.05147888511419296,
0.05457618832588196,
-0.0701843649148941,
0.04064922034740448,
0.034569162875413895,
0.09216020256280899,
-0.019903583452105522,
0.018228385597467422,
0.009555417113006115,
0.055206116288900375,
-0.04... | |
574f471a-ff39-54fb-882a-9b4edc6a6aa0 | /course/essentials/day-3/hybrid-search/ | Hybrid Search and the Universal Query API | Reciprocal Rank Fusion (RRF) > The RRF Formula | course | false | Hybrid Search and the Universal Query API > Reciprocal Rank Fusion (RRF) > The RRF Formula
$$\text{RRF\_score}(d) = \sum_{i=1}^{n} \frac{1}{k + \text{rank}_i(d)}$$
Where:
- `d` is the document
- `k` is a constant (typically 60)
- `n` is the number of retrieval methods
- `rank_i(d)` is the rank of document `d` in the ... | [
0.04532036930322647,
-0.010665769688785076,
-0.06402262300252914,
-0.06030213087797165,
0.06632627546787262,
-0.0275430865585804,
0.01712568663060665,
0.051233530044555664,
0.09991200268268585,
-0.02532496117055416,
0.002338458551093936,
0.025641055777668953,
0.05865718051791191,
0.0011840... |
5759c93f-f0db-5d09-8c83-672c3f0f1586 | /course/essentials/day-5/universal-query-demo/ | Demo: Universal Query for Hybrid Retrieval | Step 2: The Universal Query in Action > Define Global Filter with Automatic Propagation | course | true | Demo: Universal Query for Hybrid Retrieval > Step 2: The Universal Query in Action > Define Global Filter with Automatic Propagation
Now we define quality constraints that will apply throughout our entire search pipeline:
```python
# Define global filter - this will be propagated to ALL prefetch stages
global_filter ... | [
0.03795531392097473,
-0.018633201718330383,
-0.05127839371562004,
-0.03037981688976288,
0.07940199971199036,
-0.022315645590424538,
0.041998278349637985,
0.030119171366095543,
0.09941430389881134,
-0.015308263711631298,
0.03918751701712608,
0.006116415373980999,
0.08461432158946991,
-0.019... |
576e4596-4fbf-5269-81df-efa982fba707 | /documentation/manage-data/collections/ | Collections | Collection Info > Indexing Vectors in HNSW | documentation | false | Collections > Collection Info > Indexing Vectors in HNSW
In some cases, you might be surprised the value of `indexed_vectors_count` is lower than you expected. This is an intended behaviour and
depends on the [optimizer configuration](/documentation/ops-optimization/optimizer/). A new index segment is built if the siz... | [
0.020887210965156555,
-0.046662818640470505,
-0.037446025758981705,
-0.028421903029084206,
0.015332880429923534,
-0.03899379447102547,
0.0261997003108263,
0.061944492161273956,
0.05708721652626991,
-0.03475965932011604,
0.012603945098817348,
0.06495947390794754,
0.04248492419719696,
-0.019... |
5787ff98-2ef0-5612-b4eb-47e065c3054e | /documentation/search/explore/ | Explore | Recommendation API | documentation | true | Explore > Recommendation API
In addition to the regular search, Qdrant also allows you to search based on multiple positive and negative examples. The API is called ***recommend***, and the examples can be point IDs, so that you can leverage the already encoded objects; and, as of v1.6, you can also use raw vectors as... | [
0.03216163441538811,
-0.06848322600126266,
-0.03923158347606659,
-0.04368527978658676,
0.045587874948978424,
-0.00798560306429863,
0.07025101780891418,
0.0484350360929966,
0.06899809837341309,
-0.04821781814098358,
0.046953171491622925,
0.02942711114883423,
0.07219459116458893,
-0.04375390... |
57cf2869-8f30-5a40-a3f7-3f8815518819 | /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 |
|-----------------------------------... | [
0.0017666385974735022,
-0.0050475397147238255,
-0.0320868194103241,
-0.01349205058068037,
0.051925577223300934,
-0.031240319833159447,
0.05011551454663277,
0.08140457421541214,
0.04758597910404205,
-0.010034509003162384,
0.013440142385661602,
-0.01742064207792282,
0.03912544250488281,
-0.0... |
57ddcfbc-1bb7-5992-b074-0fbae78341a2 | /documentation/fastembed/fastembed-postprocessing/ | Multi-Vector Postprocessing | Using MUVERA in FastEmbed > Process with MUVERA | documentation | true | Multi-Vector Postprocessing > Using MUVERA in FastEmbed > Process with MUVERA
Now, transform the multi-vector embeddings into MUVERA's single-vector representations.
```python
muvera_embeddings = [
muvera.process_document(emb) for emb in descriptions_embeddings
]
```
Let's check the shape of a MUVERA embedding.
... | [
0.06067617982625961,
-0.05000804364681244,
-0.05257647484540939,
-0.06731801480054855,
0.04771947115659714,
-0.04170364886522293,
0.05057018622756004,
0.04884099215269089,
0.05831088125705719,
0.01878388784825802,
-0.0013567381538450718,
0.07973187416791916,
0.01772719994187355,
-0.0215649... |
57e08d42-fd97-562e-9357-b5e0132c4b3d | /documentation/frameworks/cognee/ | Cognee | A Minimal Setup | documentation | true | Cognee > A Minimal Setup
The below example comes from Cognee-community repo and mirrors the structure many teams start with: SQLite for small relational metadata, Qdrant for vectors, and Kùzu for the graph. The register import is enough for Cognee to discover and use the Qdrant adapter.
```python
import asyncio, os, ... | [
0.05029282718896866,
-0.05196603760123253,
-0.04770537093281746,
-0.015997927635908127,
0.06975334137678146,
-0.04928198456764221,
0.0771477073431015,
0.055200837552547455,
0.005520244129002094,
-0.01658731698989868,
0.02518882416188717,
0.018941299989819527,
0.05685734376311302,
-0.020027... |
5818e6f6-c5ed-5a66-a409-30ec28610eb5 | /documentation/security/ | Security | Secure Your Instance | documentation | false | Security > Secure Your Instance
Self-hosted open source deployments are not secure by default and are not production-ready. Qdrant Cloud deployments are always secure and production-ready.
By default, all self-deployed Qdrant instances are not secure. They are open to
all network interfaces and do not have any kind o... | [
0.031871747225522995,
-0.0644766315817833,
-0.012764356099069118,
-0.03082774206995964,
0.0385403074324131,
-0.044320084154605865,
0.0647127702832222,
0.05009350925683975,
0.07812266796827316,
-0.003410967532545328,
0.010696347802877426,
0.01037792582064867,
0.039504218846559525,
-0.064922... |
581e54c3-3627-517d-a86b-ff2e34e0279a | /documentation/frameworks/mastra/ | Mastra | Usage | documentation | true | Mastra > Usage
```typescript
import { QdrantVector } from "@mastra/rag";
const qdrant = new QdrantVector({
url: "https://xyz-example.eu-central.aws.cloud.qdrant.io:6333"
apiKey: "<YOUR_API_KEY>",
https: true
});
``` | [
0.049750421196222305,
-0.004723278805613518,
-0.03939460963010788,
-0.07055375725030899,
0.0352054201066494,
-0.011072334833443165,
0.045424528419971466,
0.004859456792473793,
0.013400174677371979,
-0.020480915904045105,
0.03459616377949715,
0.004663161933422089,
0.037616901099681854,
-0.0... |
58230cd4-0373-5f8d-8d89-9fe66f0fd9b6 | /documentation/embeddings/jina-embeddings/ | Jina AI | Embed Queries | documentation | true | Jina AI > Embed Queries
v5 models have task-specific adapters. Embed queries with `retrieval.query` and documents with `retrieval.passage`, so the model encodes a short question and a longer passage that answers it close together. The default task is `text-matching`, which is for symmetric similarity, so always set th... | [
0.05027599632740021,
-0.043010178953409195,
-0.031015878543257713,
-0.04639656841754913,
0.07404495775699615,
-0.025919660925865173,
0.04822338744997978,
0.016385043039917946,
0.05823102593421936,
0.00023006065748631954,
0.05987357348203659,
0.004582069348543882,
0.02279902622103691,
-0.01... |
582a163c-6009-5b6d-acdd-9f9bfbaf51e4 | /articles/late-interaction-models/ | Late Interaction Retrieval with Dense Token Embeddings | Limitations and Further Reading | article | false | Late Interaction Retrieval with Dense Token Embeddings > Limitations and Further Reading
These results come from small BEIR datasets evaluated in exact search mode, measured in 2024. We did not measure retrieval latency, and the gains varied by dataset.
If you want to dig into the storage and speed costs of multi-vec... | [
0.06995413452386856,
-0.03048679232597351,
-0.041889600455760956,
-0.061153579503297806,
0.05765330418944359,
-0.03056948073208332,
0.050143394619226456,
0.025201993063092232,
0.07798213511705399,
-0.027374401688575745,
0.02296328917145729,
0.02247416041791439,
0.035435374826192856,
-0.015... |
582f8d97-c91c-55f9-9982-cab0a4ec7c60 | /course/essentials/day-2/what-is-hnsw/ | HNSW Indexing Fundamentals | course | false | HNSW Indexing Fundamentals
Day 2
At this point, you've learned how vector search retrieves the nearest vectors to a query using cosine similarity, dot product, or Euclidean distance. How does this work at scale? | [
0.05381760746240616,
-0.057498589158058167,
-0.06371945887804031,
-0.026822630316019058,
0.04909072816371918,
-0.03911001607775688,
0.0341508686542511,
0.033571451902389526,
0.0717168003320694,
-0.0044997334480285645,
0.033013492822647095,
0.058777470141649246,
0.03364544361829758,
-0.0101... | |
582fdbda-73eb-502d-a2b4-e39515a74876 | /course/essentials/day-3/hybrid-search/ | Hybrid Search and the Universal Query API | Fusion: Combining Search Signals | course | false | Hybrid Search and the Universal Query API > Fusion: Combining Search Signals
Each retrieval method produces a numerical score assigned to each document it selects. These scores can't be directly compared to each other between different methods, so there is no clear way to mix candidates.
**The scoring problem:**
- Yo... | [
0.040825653821229935,
-0.04811255633831024,
-0.08868534117937088,
-0.04532766342163086,
0.07953231781721115,
-0.035519421100616455,
0.06162885203957558,
0.019146060571074486,
0.06490504741668701,
-0.011620355769991875,
0.056393083184957504,
0.02175854705274105,
0.07988893985748291,
0.00454... |
585377b7-79ab-598e-bc4b-b869f97ff72c | /documentation/frameworks/ | Frameworks | Framework Integrations | documentation | false | Frameworks > Framework Integrations
| Framework | Description |
| ------------------------------------------------------------- | --------------------------------... | [
0.036464203149080276,
-0.011678170412778854,
-0.06758460402488708,
-0.028955306857824326,
0.10163121670484543,
-0.045839060097932816,
0.0540132112801075,
-0.00047526208800263703,
0.020925048738718033,
-0.07129260897636414,
0.027837445959448814,
-0.0071915267035365105,
0.02262766659259796,
... |
585f639c-f546-5ce8-9c9d-f163a10b91a8 | /documentation/tutorials-operations/scaling-and-monitoring-pods/ | Scaling and Monitoring Hybrid Cloud | Check What Your Pod Can Actually Use | documentation | true | Scaling and Monitoring Hybrid Cloud > Check What Your Pod Can Actually Use
Once you've scaled, it's tempting to assume Qdrant can use the full amount of resources you just set, but it can't. Hybrid Cloud reserves part of each pod's CPU and memory for the operating system, Kubernetes, and other system components, and t... | [
0.03458075225353241,
0.028674039989709854,
-0.035472407937049866,
-0.03784554824233055,
0.07647497951984406,
-0.02319490909576416,
0.05928358435630798,
0.04998230189085007,
0.07564273476600647,
-0.057554446160793304,
0.04031515493988991,
-0.014687851071357727,
0.05423283204436302,
-0.02305... |
586483b1-9f0d-50cc-be09-185e251c0144 | /documentation/search/text-search/full-text-search/ | Full-Text Search | BM25 > Configuring BM25 Parameters | documentation | true | Full-Text Search > BM25 > Configuring BM25 Parameters
The BM25 [ranking function](https://en.wikipedia.org/wiki/Okapi_BM25#The_ranking_function) has two parameters that you can set per query:
- `k`. Controls term frequency saturation. Higher values increase the influence of term frequency. Must be 0 or greater. Defau... | [
0.009364691562950611,
-0.019157782196998596,
-0.016610193997621536,
-0.04633002728223801,
0.026791084557771683,
-0.0469442717730999,
0.047828540205955505,
0.07667911052703857,
0.09421133995056152,
0.008761657401919365,
0.012851410545408726,
-0.009363577701151371,
0.05972850322723389,
-0.02... |
587b1889-c126-5f3e-8a31-1448970ac378 | /documentation/search/text-search/sparse-vector-search/ | Sparse Vector Search | BM25 > BM25 Text Processing > Stemming and Stopwords | documentation | true | Sparse Vector Search > BM25 > BM25 Text Processing > Stemming and Stopwords
To configure stemming and stopword removal, use the following options:
- `language`: sets the language for stemming and stopword removal. Defaults to `english`.
- `stemmer`: defaults to stemming for `language` (if set), but can be configured ... | [
0.042573172599077225,
-0.04684754088521004,
0.027467695996165276,
-0.03320445865392685,
0.029306864365935326,
-0.05070165917277336,
0.03910469263792038,
0.046699896454811096,
0.033086687326431274,
-0.029618972912430763,
0.022428058087825775,
0.05429678037762642,
0.03908265382051468,
-0.008... |
58842b5c-3fae-5b0f-bafe-08185dbea39a | /course/multi-vector-search/module-3/final-project/ | Final Project: Build Your Own Multi-Vector Search System | Your Mission | course | false | Final Project: Build Your Own Multi-Vector Search System > Your Mission
Module 3
---
It's time to bring together everything you've learned about multi-vector search, late interaction models, and production optimization. You'll build a sophisticated document retrieval system that leverages late interaction's token-le... | [
0.03568876534700394,
-0.03201131150126457,
-0.02582632377743721,
-0.029257001355290413,
0.1071920320391655,
-0.021969078108668327,
0.05072912201285362,
0.036602165549993515,
0.067327581346035,
-0.019216490909457207,
-0.0026780343614518642,
0.013243453577160835,
0.06048902869224548,
-0.0395... |
58983b6c-789a-567d-af7d-bbbae8df96c7 | /documentation/cloud/create-cluster/ | Create a Cluster | Create a Cluster | documentation | false | Create a Cluster > Create a Cluster
1. Choose your data center region or Hybrid Cloud environment.
1. Configure RAM for each node.
> For more information, see our [Capacity Planning](/documentation/capacity-planning/) guidance.
1. Choose the number of vCPUs and GPUs per node. If you add more
RAM, the menu pro... | [
-0.004884249530732632,
-0.023612795397639275,
-0.07667619735002518,
-0.041170280426740646,
0.07570601254701614,
-0.018989665433764458,
0.03372148796916008,
0.03744615986943245,
0.08696117252111435,
-0.03146364539861679,
0.018173931166529655,
0.000825161230750382,
0.051856014877557755,
-0.0... |
589eaffb-438e-5df5-a995-db066231f3b2 | /documentation/tutorials-build-essentials/data-ingestion-beginners/ | S3 Ingestion | Architecture | documentation | false | S3 Ingestion > Architecture
Each product has its own folder in the bucket, holding a text review and an image. A Python script lists the folders with `boto3`, downloads each review and image, and hands them to `qdrant-client`. The client embeds the text and the image locally with FastEmbed and upserts one point per pr... | [
0.01339996512979269,
-0.04943687096238136,
-0.05658735707402229,
-0.03607400506734848,
0.04330870136618614,
-0.025420449674129486,
0.055041711777448654,
0.009152251295745373,
0.04570259526371956,
-0.029191436246037483,
0.01596062071621418,
0.019088178873062134,
0.049612440168857574,
-0.031... |
58b0aa6b-7d7f-5599-aa9e-b71e2430e76c | /documentation/manage-data/indexing/ | Indexing | Full-Text Index > Tokenizers > Multilingual | documentation | false | Indexing > Full-Text Index > Tokenizers > Multilingual
The `multilingual` tokenizer is based on multiple packages like [charabia](https://github.com/meilisearch/charabia) and [vaporetto](https://github.com/daac-tools/vaporetto) to deliver fast and accurate tokenization for a large variety of languages. It allows prope... | [
0.014687098562717438,
-0.007831205613911152,
-0.05457094684243202,
-0.08270280063152313,
0.06825055927038193,
-0.04994361847639084,
0.09393765777349472,
0.02288850210607052,
0.06093311309814453,
-0.021000957116484642,
0.052565254271030426,
0.041045915335416794,
0.04479752108454704,
0.00295... |
58bdb723-e03b-58f1-94cc-66ea2ce99d10 | /course/beginners/module-2/further-reading/ | Further Reading | course | false | Further Reading
Module 2
- [Distance Metrics](/course/essentials/day-1/distance-metrics/) A closer look at cosine similarity, dot product, Euclidean, and Manhattan, and when each one fits.
- [What Is HNSW](/course/essentials/day-2/what-is-hnsw/) How the graph index is built and tuned, once you have real searches to m... | [
0.03939361125230789,
-0.017152387648820877,
-0.06326956301927567,
-0.021897025406360626,
0.060428112745285034,
-0.03013475425541401,
0.07447747141122818,
0.06825601309537888,
0.08133213967084885,
-0.04553719609975815,
0.034958530217409134,
0.06658216565847397,
0.024920517578721046,
-0.0122... | |
58bdead1-ab1c-570b-88d4-789a9e10a2e5 | /documentation/ops-optimization/optimizer/ | Optimizer | Prevent Reads from Large Unindexed Segments > Effect on wait=true | documentation | false | Optimizer > Prevent Reads from Large Unindexed Segments > Effect on wait=true
Because the update worker must finish indexing before continuing to consume the queue, a blocked `wait=true` request also delays all subsequent updates that use `wait=true`. Updates with `wait=false` are written to the write-ahead log immedi... | [
0.01642366498708725,
-0.024185461923480034,
-0.06871279329061508,
-0.029364902526140213,
0.05950259044766426,
-0.02595473825931549,
0.018296781927347183,
0.0020379750058054924,
0.05651118606328964,
-0.021301930770277977,
0.020184554159641266,
0.040028832852840424,
0.08953878283500671,
-0.0... |
58c8fb0e-e8c6-55d3-b65e-9ec9ab7023c6 | /documentation/migrate-to-qdrant/from-mongodb/ | From MongoDB | Concept Mapping | documentation | false | From MongoDB > Concept Mapping
| MongoDB | Qdrant | Notes |
| :--- | :--- | :--- |
| Collection | Collection | One-to-one mapping |
| Document | Point | Each document becomes a point |
| Vector field | Vector | Named vectors are preserved |
| Non-vector fields | Payload | Direct mapping |
| `_id` (ObjectId or string) ... | [
0.06622789800167084,
-0.05290968716144562,
-0.04711202159523964,
-0.05248737335205078,
0.0725276991724968,
-0.014108596369624138,
0.07668659090995789,
0.07021206617355347,
0.061504945158958435,
-0.056093379855155945,
0.03444793075323105,
0.013424532487988472,
0.041035402566194534,
-0.01512... |
58e9edd7-383d-54d9-9356-5558c1ad47c6 | /documentation/private-cloud/qdrant-cluster-management/ | Managing a Cluster | Resharding | documentation | false | Managing a Cluster > Resharding
In Qdrant Cloud, you can change the number of shards in your existing collections without having to recreate the collection from scratch. This feature is called resharding and allows you to scale your collections up or down as needed. For more details see [Resharding](/documentation/clo... | [
0.04027390107512474,
-0.018067307770252228,
-0.0463932566344738,
-0.07981886714696884,
0.06535894423723221,
-0.05930311977863312,
0.027472257614135742,
0.036943335086107254,
0.06019533425569534,
-0.041642095893621445,
0.012234843336045742,
0.0026011234149336815,
0.05321747064590454,
-0.000... |
59172bcd-cb58-5c3f-aa9b-ba44728039ae | /course/multi-vector-search/module-2/visual-interpretability/ | Visual Interpretability of ColPali | course | false | Visual Interpretability of ColPali
Module 2
**Why did this document match my query?** Unlike traditional black-box embedding models that produce a single opaque vector, ColPali's multi-vector architecture offers something remarkable: you can see exactly where the model "looks" when matching a query to a document.
Th... | [
0.025114798918366432,
-0.06996416300535202,
-0.060756027698516846,
-0.04056414216756821,
0.06571225076913834,
-0.007436138112097979,
0.0907926857471466,
0.046109434217214584,
0.011678517796099186,
-0.000332733296090737,
0.029480552300810814,
-0.007427075412124395,
0.0792265459895134,
-0.01... | |
5939a337-fa42-5884-b0a5-1866f80d3c3b | /documentation/search-tuning/when-a-reranker-is-worth-it/ | When Is a Reranker Worth It? | What to Tune Next | documentation | false | When Is a Reranker Worth It? > What to Tune Next
After a win, the work moves to throughput, where the candidate count you can serve decides how much of the gain survives. After a loss, the gap is still there and the candidates are what to change, so revisit retrieval and [candidate depth](/documentation/search-tuning/... | [
0.01601286418735981,
-0.06025926023721695,
-0.010894005186855793,
-0.03874628245830536,
0.041626572608947754,
-0.0157060157507658,
0.02107873745262623,
0.0035379277542233467,
0.09430424124002457,
-0.002198112430050969,
0.025693655014038086,
0.00871354527771473,
0.06367270648479462,
0.00639... |
593aa0e0-2e77-5e70-8be9-612c89a80489 | /articles/what-are-embeddings/ | Vector Embeddings Explained: How They Work in ML & Search | Create a Neural Search Service | article | false | Vector Embeddings Explained: How They Work in ML & Search > Create a Neural Search Service
Now that you’re familiar with the core concepts around vector embeddings, how about start building your own [Neural Search Service](/documentation/tutorials-develop/neural-search/)?
This tutorial guides you through a practical ... | [
0.05451592803001404,
-0.05639413744211197,
-0.05119997262954712,
-0.001200554776005447,
0.03962922468781471,
-0.037293266505002975,
0.09434269368648529,
0.06978639960289001,
0.06031188741326332,
-0.004114037845283747,
0.020554441958665848,
0.035902030766010284,
0.02295936457812786,
-0.0541... |
59425fd3-94f5-51f3-b470-b121f29eb0c1 | /documentation/tutorials-build-essentials/video-anomaly-edge-part-3/ | Video Anomaly Detection Part 3: Scoring, Governance, and Deployment | Results > How the scores are computed | documentation | false | Video Anomaly Detection Part 3: Scoring, Governance, and Deployment > Results > How the scores are computed
The numbers below depend on choices that are easy to miss, so here is exactly what is measured:
- **Baseline.** Only normal training footage, cut into 10-second clips and embedded. No anomalous clip is used, wh... | [
0.04603250324726105,
-0.04586293175816536,
-0.09591551125049591,
-0.04100920632481575,
0.05010763555765152,
0.0058035049587488174,
0.030635764822363853,
0.021291173994541168,
0.0900755301117897,
-0.06388459354639053,
0.023877114057540894,
0.008030022494494915,
0.07509809732437134,
-0.03949... |
595160df-4899-5b1b-85a9-a6840ae3979b | /documentation/search-patterns/choose-embedding-model/ | How to Choose an Embedding Model: Evaluation & Tradeoffs | Evaluation: The Holy Grail of Vector Search > Know the Language Your Model Speaks | documentation | false | How to Choose an Embedding Model: Evaluation & Tradeoffs > Evaluation: The Holy Grail of Vector Search > Know the Language Your Model Speaks
Normalization can also erase differences in your text. The `all-MiniLM-L6-v2` tokenizer lowercases text and strips accents, so "crème brûlée" and "creme brulee" reach the model a... | [
0.06794317066669464,
-0.04343622177839279,
-0.0369853675365448,
-0.04102292284369469,
0.06793136149644852,
-0.021415654569864273,
0.05893344059586525,
0.0588354654610157,
0.035147350281476974,
-0.0251690112054348,
0.04350383207201958,
0.0025034502614289522,
0.05925056338310242,
-0.03188996... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.