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 |
|---|---|---|---|---|---|---|---|
f0a17a9c-82e8-58ea-96f6-8a7c33be4cb1 | /course/essentials/day-1/embedding-models/ | Points, Vectors and Payloads | Key Takeaways | course | false | Points, Vectors and Payloads > Key Takeaways
Understanding Qdrant's data model prepares you for building sophisticated search applications. Points combine unique IDs, vectors, and metadata into a flexible foundation. Multiple vector types (dense, sparse, multivector) support different use cases, while named vectors en... | [
0.032675642520189285,
-0.038956742733716965,
-0.0335908941924572,
-0.021601641550660133,
0.02275686338543892,
-0.07387164980173111,
0.10253514349460602,
0.06992214173078537,
0.08786947280168533,
-0.026719363406300545,
0.013921985402703285,
0.0330764576792717,
0.014021575450897217,
0.002027... |
f0f5b4db-cb86-51d6-ab34-9de4bf2d2a55 | /course/essentials/day-6/congratulations/ | Course Completion and Next Steps | Your Portfolio Piece | course | false | Course Completion and Next Steps > Your Portfolio Piece
Your final project serves as a comprehensive demonstration of modern vector search capabilities. It showcases not just technical implementation skills, but also the systems thinking and evaluation rigor that distinguishes senior engineers.
The combination of hyb... | [
0.012582159601151943,
-0.05360877513885498,
-0.029414312914013863,
-0.040931958705186844,
0.06508403271436691,
-0.046155329793691635,
0.0706402137875557,
0.029517028480768204,
0.06181316822767258,
-0.02245379611849785,
0.020357737317681313,
0.039884861558675766,
0.06597912311553955,
-0.029... |
f0f874a9-db2d-5936-bc49-d1beacb77184 | /articles/immutable-data-structures/ | Immutable Data Structures in Qdrant | Data Structures 101 > Mutability | article | true | Immutable Data Structures in Qdrant > Data Structures 101 > Mutability
One of the most significant challenges when working with data structures is ensuring **mutability — the ability to change the data structure after it’s created**, particularly with fast update operations.
Let’s consider a simple example: we want t... | [
0.05161849036812782,
-0.03654012456536293,
-0.07218228280544281,
-0.026204217225313187,
0.0837356299161911,
-0.05385074019432068,
0.05385540798306465,
-0.0013397083384916186,
0.06368996202945709,
-0.012673449702560902,
0.015306556597352028,
0.01677645556628704,
0.06553708016872406,
-0.0073... |
f0fa0d3a-c779-5185-87f0-28fe9a4deb60 | /documentation/tutorials-build-essentials/semantic-cache/ | Semantic Caching for RAG | Serve from the Cache | documentation | true | Semantic Caching for RAG > Serve from the Cache
A hit serves the stored answer without an LLM call. A miss retrieves policy documents from Qdrant, generates an answer, and stores the question and answer.
Let's put all of this together.
```python
THRESHOLD = 0.98
class Served(TypedDict):
answer: str
hit: boo... | [
0.025781527161598206,
-0.031106723472476006,
-0.05549116060137749,
-0.022818490862846375,
0.055711667984724045,
0.003016681643202901,
0.03383876010775566,
0.012237990275025368,
0.0561273917555809,
-0.04619291424751282,
0.026724286377429962,
0.027311241254210472,
0.05918125435709953,
-0.013... |
f0fe50c6-c433-531e-b386-e2abb1215330 | /documentation/search-patterns/vector-search-filtering/ | Filtering in Vector Search: When to Use What | Further Reading | documentation | false | Filtering in Vector Search: When to Use What > Further Reading
- [Filtering documentation](/documentation/search/filtering/): every filter condition and clause, with examples for each client.
- [Payload indexes](/documentation/manage-data/indexing/#payload-index): index types, their parameters, and the tenant and prin... | [
0.03657187148928642,
-0.05172701179981232,
-0.06196283549070358,
-0.018828878179192543,
0.07176855206489563,
-0.01979404129087925,
0.04082288593053818,
0.024533774703741074,
0.07407034933567047,
-0.017223089933395386,
0.02549651265144348,
0.05867490544915199,
0.04407168924808502,
-0.032312... |
f113696e-f54e-587e-9159-ab9d51b95de4 | /documentation/manage-data/indexing/ | Indexing | Payload Index > Parameterized Index | documentation | false | Indexing > Payload Index > Parameterized Index
Beyond selecting the field type, you can set parameters on a payload index to fine-tune how it is stored and which filtering conditions it can serve. The available parameters depend on the field type, and are described in the subsections below. | [
0.02551053836941719,
-0.03300224989652634,
-0.03272474929690361,
-0.0396902896463871,
0.017544426023960114,
-0.03930557891726494,
0.006077250931411982,
0.039130669087171555,
0.06907951086759567,
-0.023366734385490417,
0.06043855473399162,
0.038831379264593124,
0.0602010041475296,
-0.012553... |
f12d5958-20fc-51f8-9d8c-772ae755bab0 | /documentation/manage-data/indexing/ | Indexing | Sparse Vector Index | documentation | true | Indexing > Sparse Vector Index
*Available as of v1.7.0*
Sparse vectors in Qdrant are indexed with a special data structure, which is optimized for vectors that have a high proportion of zeroes. In some ways, this indexing method is similar to the inverted index, which is used in text search engines.
- A sparse vecto... | [
0.06014734506607056,
-0.060065798461437225,
-0.045554131269454956,
-0.015179996378719807,
0.07730275392532349,
-0.037086255848407745,
0.04396035894751549,
0.007638740353286266,
0.09355143457651138,
-0.02708144299685955,
-0.009444359689950943,
0.024021342396736145,
0.02665645442903042,
-0.0... |
f156e646-12c6-5b4c-a063-97f14d323463 | /documentation/frameworks/dynamiq/ | Dynamiq | documentation | true | Dynamiq
Dynamiq is your all-in-one Gen AI framework, designed to streamline the development of AI-powered applications. Dynamiq specializes in orchestrating retrieval-augmented generation (RAG) and large language model (LLM) agents.
Qdrant is a vector database available in Dynamiq, capable of serving multiple roles. ... | [
0.02171550877392292,
-0.0496685616672039,
-0.05639201030135155,
-0.034121282398700714,
0.06549396365880966,
-0.0347907654941082,
0.06827715784311295,
0.04100210964679718,
0.0355721116065979,
-0.05994802713394165,
0.02675952948629856,
0.03318218141794205,
0.02837490476667881,
-0.01930982246... | |
f15ab83a-c354-5930-bd00-84fc31417201 | /course/multi-vector-search/module-3/pooling-techniques/ | Pooling Techniques | What's Next | course | false | Pooling Techniques > What's Next
This lesson covered two complementary strategies for reducing the number of vectors per document:
- **Row/column pooling**: Exploits spatial structure in image embeddings for a fixed reduction (32x for ColPali)
- **Hierarchical pooling**: Content-adaptive clustering that works for any... | [
0.0490039587020874,
-0.04065857455134392,
-0.0478132963180542,
-0.038478683680295944,
0.07222282141447067,
-0.03272503986954689,
0.020527079701423645,
0.03198773041367531,
0.0771862342953682,
-0.0003321472031529993,
0.029078003019094467,
0.030549021437764168,
0.0419318750500679,
-0.0257901... |
f17e1802-c4fe-5a66-9065-0753ba738791 | /documentation/examples/rag-chatbot-vultr-dspy-ollama/ | Private RAG Information Extraction Engine | Implementing the information extraction pipeline > Application logic | documentation | true | Private RAG Information Extraction Engine > Implementing the information extraction pipeline > Application logic
There is a concept of signatures which defines input and output formats of the pipeline. We are going to define a simple
signature for the event:
```python
class Event(dspy.Signature):
description = ds... | [
0.02439192868769169,
-0.04389326646924019,
-0.05387960001826286,
-0.009651090949773788,
0.0723586156964302,
-0.06353536993265152,
0.056175343692302704,
0.056802332401275635,
0.033135347068309784,
-0.032473959028720856,
-0.0034933097194880247,
0.0436239093542099,
0.06345273554325104,
-0.044... |
f1896767-b508-5c91-b04c-1f835f584b03 | /documentation/overview/ | Overview | Deployments | documentation | false | Overview > Deployments
Qdrant supports multiple deployment models to match different infrastructure and operational needs. The right option depends on your security constraints and operational model: Qdrant-managed infrastructure ([Managed Cloud](/documentation/cloud/)), shared responsibility with your own clusters ([... | [
-0.0007359945448115468,
-0.026043839752674103,
-0.05514422059059143,
-0.04343070834875107,
0.056514907628297806,
-0.02595224604010582,
0.11887657642364502,
0.024446334689855576,
0.05944100394845009,
-0.010526832193136215,
0.036666762083768845,
-0.014280589297413826,
0.02503969706594944,
-0... |
f19a8d79-fd70-59b9-bc14-2d144821c67a | /articles/late-interaction-models/ | Late Interaction Retrieval with Dense Token Embeddings | Practical Application: Enhancing Retrieval with Dense Models | article | false | Late Interaction Retrieval with Dense Token Embeddings > Practical Application: Enhancing Retrieval with Dense Models
If you’re using one of the sentence transformer models, the output token embeddings are calculated by default. While a single vector representation is more efficient in terms of storage and computation... | [
0.05876336991786957,
-0.04469319432973862,
-0.033036261796951294,
-0.03872830420732498,
0.05015910416841507,
-0.05886811017990112,
0.06107807531952858,
0.0773204043507576,
0.07544700801372528,
-0.02731166034936905,
0.0052280789241194725,
0.043893203139305115,
0.024401362985372543,
-0.02804... |
f1b2069c-c1fa-5b22-ad96-8416f52880c3 | /documentation/production-operations/managed-cloud-prometheus/ | Managed Cloud Prometheus Monitoring | Step 1: Install Prometheus and Grafana | documentation | true | Managed Cloud Prometheus Monitoring > Step 1: Install Prometheus and Grafana
If you haven't installed Prometheus and Grafana yet, you can use the [kube-prometheus-stack](https://artifacthub.io/packages/helm/prometheus-community/kube-prometheus-stack) Helm chart to deploy them in your Kubernetes cluster.
A minimal exa... | [
0.028886642307043076,
-0.04745146632194519,
-0.027315475046634674,
-0.025270745158195496,
0.05349106341600418,
-0.02887134812772274,
0.035932835191488266,
0.048446718603372574,
0.03982880339026451,
-0.03230952098965645,
0.05616673827171326,
0.02253728359937668,
0.06364742666482925,
-0.0113... |
f1c157cb-7840-5b36-93eb-6e94b799a5ff | /documentation/search/text-search/ | Text Search | documentation | false | Text Search
Qdrant is a vector search engine, making it a great tool for [semantic search](#semantic-search). However, Qdrant's capabilities go beyond just vector search. It also supports a range of lexical search features, including filtering on text fields and full-text search using popular algorithms like BM25. | [
0.04032251983880997,
-0.0545911006629467,
-0.051065199077129364,
-0.025694681331515312,
0.03745659813284874,
-0.015761593356728554,
0.10364201664924622,
-0.009082400240004063,
0.0534808374941349,
-0.008848322555422783,
0.05513190105557442,
0.01910063810646534,
0.06050596758723259,
-0.03574... | |
f1d2ad68-faf0-548d-a25a-79c4edb49ca4 | /documentation/cloud/configure-cluster/ | Configure Clusters | Client IP Restrictions | documentation | false | Configure Clusters > Client IP Restrictions
If configured, only the chosen IP ranges will be allowed to access the cluster. This is useful for securing your cluster and ensuring that only clients coming from trusted networks can connect to it.
 | [
0.004245774820446968,
-0.01915489137172699,
-0.03559889644384384,
-0.02672276645898819,
0.03242575749754906,
-0.07114771753549576,
0.02140231616795063,
0.08423887193202972,
0.03589164465665817,
0.022983500733971596,
0.005393712781369686,
0.017260875552892685,
0.09426596015691757,
-0.019594... |
f1dcabc7-1d2e-5ef5-9060-9906a8500265 | /documentation/edge/edge-api/reading-data/ | Reading Data | query_groups | documentation | true | Reading Data > query_groups
*Rust only*
Groups query results by a payload field, returning a bounded number of hits per distinct value.
```rust
fn query_groups(&self, request: GroupRequest) -> OperationResult<Vec<Group>>
```
**Returns** a list of `Group`, each carrying the group's `key` and its `hits`.
| Parameter... | [
0.021757936105132103,
-0.03717603161931038,
-0.03590703010559082,
-0.017231430858373642,
0.08044841140508652,
-0.009438804350793362,
0.04936916008591652,
0.02323569543659687,
0.0792188048362732,
-0.026983926072716713,
0.06019263342022896,
0.01879165694117546,
0.03578457981348038,
-0.023631... |
f1e26c27-a69b-54d7-8da6-b8c8d6fd8248 | /course/essentials/day-0/qdrant-cloud/ | Qdrant Setup | Access the Web UI > What you can do in the Web UI > Main Navigation | course | false | Qdrant Setup > Access the Web UI > What you can do in the Web UI > Main Navigation
**Console**: Run REST calls in the browser. Test endpoints, inspect responses, and debug queries without writing code. Handy for exploring the full API.
**Collections**: See and manage all collections. Create collections, upload snapsh... | [
0.036031465977430344,
-0.04655662924051285,
-0.039763569831848145,
-0.020758626982569695,
0.04190967604517937,
-0.05004260689020157,
0.052497223019599915,
0.023375483229756355,
0.04417762532830238,
-0.03282729536294937,
0.047526001930236816,
0.0449877493083477,
0.046538129448890686,
-0.046... |
f20984ae-01fc-516a-afb3-5baa12f4145d | /documentation/embeddings/ollama/ | Ollama | documentation | false | Ollama
[Ollama](https://ollama.com) downloads open embedding models and serves them from your own machine through a local HTTP API on port 11434. Your application sends text to Ollama's `/api/embed` endpoint and stores the returned vectors in Qdrant, so the text never leaves your infrastructure. | [
0.05434884876012802,
-0.0559096522629261,
-0.02022024802863598,
-0.05906428024172783,
0.06661005318164825,
-0.02486947737634182,
0.07172904163599014,
0.0682820975780487,
0.028401963412761688,
-0.008807439357042313,
0.0028988870326429605,
0.004339967854321003,
0.019571904093027115,
-0.02489... | |
f214f34a-db32-536c-8c96-f9ab0fdaa238 | /documentation/migration-recovery-options/ | Migration and Recovery | Comparison | documentation | false | Migration and Recovery > Comparison
The following tables compare the three options across key dimensions to help you choose the right approach for your migration or recovery needs. | [
0.03463596850633621,
-0.04378724470734596,
-0.03149464726448059,
-0.07523661851882935,
0.038735464215278625,
0.01323399692773819,
0.05546486750245094,
0.0380440428853035,
0.060610897839069366,
-0.006671177688986063,
0.012123235501348972,
0.037218380719423294,
0.06917298585176468,
-0.006486... |
f21ecdb6-fbfa-5c4c-bc0b-7f44819ea439 | /course/essentials/day-1/chunking-strategies/ | Text Chunking Strategies | What's Next | course | false | Text Chunking Strategies > What's Next
Now that you understand how to structure and prepare textual data, let's put these concepts into practice. In the next section, we'll build a complete movie recommendation system that demonstrates chunking, embedding, and metadata in action.
Remember: **Qdrant doesn't make assum... | [
0.046247899532318115,
-0.04117225483059883,
-0.05010848864912987,
-0.05363065376877785,
0.03837788477540016,
-0.03632181137800217,
0.059753984212875366,
0.018644293770194054,
0.04557884484529495,
-0.015769263729453087,
0.04852364957332611,
-0.012019895017147064,
0.1000647097826004,
-0.0153... |
f237b443-c91f-5a3d-bc9c-03fbf3d52c34 | /documentation/security/ | Security | Authentication | documentation | false | Security > Authentication
By default, an open source Qdrant deployment accepts requests from anyone who can reach it. To secure your instance, enable API key authentication. On Qdrant Cloud, API key authentication is enabled by default.
Qdrant supports three types of API key:
- **[Admin API Key](#admin-api-key)**: G... | [
0.014635950326919556,
-0.03208368644118309,
-0.029225431382656097,
-0.06371522694826126,
0.04224646836519241,
-0.06353921443223953,
0.06536392867565155,
0.060558877885341644,
0.0871765986084938,
-0.017911463975906372,
0.03750227391719818,
0.031742170453071594,
0.054465677589178085,
-0.0380... |
f2471eec-2586-5fc7-8198-9a16fcdabbf5 | /documentation/migrate-to-qdrant/from-qdrant/ | From Qdrant | Gotchas | documentation | false | From Qdrant > Gotchas
- **Parallel workers and resume:** Changing `--migration.num-workers` between runs starts the migration over instead of resuming it.
- **Large messages:** If you encounter gRPC message size errors, increase `--source.max-message-size`.
## Next Steps
After migration, verify your data arrived cor... | [
0.0004377855220809579,
-0.034262675791978836,
-0.05840460583567619,
-0.014030678197741508,
0.07737720757722855,
-0.01721974089741707,
0.037086665630340576,
0.04013598710298538,
0.042363982647657394,
0.002198008121922612,
0.03712073713541031,
0.012931863777339458,
0.02107463963329792,
-0.06... |
f25e9c60-2074-59f5-a29f-5a4a9c9c7f69 | /documentation/search/search/ | Search | Grouping API > Lookup in Groups | documentation | true | Search > Grouping API > Lookup in Groups
The looked-up result appears under `lookup` in each group. Below, chunk id 1 shows up in both groups because its `document_id` payload is an array (`[200, 201]`), placing the chunk into every matching group.
```json
{
"result": {
"groups": [
{
... | [
0.0062859985046088696,
-0.05956035479903221,
-0.04888517037034035,
-0.02695498615503311,
0.1153530478477478,
-0.025249719619750977,
0.04769201949238777,
0.03300999104976654,
0.07150047272443771,
-0.0585956797003746,
0.023730795830488205,
0.01403011567890644,
0.03746747598052025,
0.00510477... |
f26f0d8b-025e-566c-a872-fc66b116f31c | /documentation/private-cloud/qdrant-cluster-management/ | Managing a Cluster | Audit Logging | documentation | true | Managing a Cluster > Audit Logging
*Available as of Qdrant v1.17.0*
Audit logging records API operations that require authentication or authorization to JSON log files on the database volume. It is disabled by default. Enable it under `spec.config.audit` on the `QdrantCluster`:
```yaml
apiVersion: qdrant.io/v1
kind:... | [
0.007104766089469194,
-0.04402259737253189,
-0.010876319371163845,
-0.04205989092588425,
0.039682406932115555,
-0.07560581713914871,
0.013849202543497086,
0.06952481716871262,
0.04936334490776062,
-0.04742670804262161,
0.006596519611775875,
0.036286864429712296,
0.04617563262581825,
-0.027... |
f2837781-b1d0-5ebc-bd4f-d9af10cb594d | /articles/modern-sparse-neural-retrieval/ | Modern Sparse Neural Retrieval: From DeepCT to SPLADE++ | Did we Solve the Vocabulary Mismatch Yet? > External Document Expansion > External Document Expansion with docT5query | article | false | Modern Sparse Neural Retrieval: From DeepCT to SPLADE++ > Did we Solve the Vocabulary Mismatch Yet? > External Document Expansion > External Document Expansion with docT5query
[`docT5query`](https://github.com/castorini/docTTTTTquery) is the most used document expansion model.
It is based on the [Text-to-Text Transfe... | [
0.055751532316207886,
-0.03989395126700401,
-0.019064946100115776,
-0.031585466116666794,
0.057063329964876175,
-0.03462487831711769,
0.06535869836807251,
0.01732754148542881,
0.0915275290608406,
-0.03269996494054794,
0.007497183512896299,
0.0051515428349375725,
-0.0027689349371939898,
-0.... |
f2acb62e-e7fe-57c3-a36c-5f701e76652b | /documentation/tutorials-build-essentials/video-anomaly-edge-part-2/ | Video Anomaly Detection Part 2: Edge-to-Cloud Pipeline | Edge-to-Cloud Escalation | documentation | false | Video Anomaly Detection Part 2: Edge-to-Cloud Pipeline > Edge-to-Cloud Escalation

When an edge device scores a clip above the escala... | [
0.033694788813591,
-0.021931827068328857,
-0.07323431223630905,
-0.06540925055742264,
0.02234533429145813,
0.01312567014247179,
0.057452745735645294,
0.05401722714304924,
0.07341717183589935,
-0.02708354778587818,
0.02215721644461155,
0.030705425888299942,
0.05899127572774887,
-0.006122223... |
f2b0ce97-fa6d-51b6-8701-494276687c47 | /documentation/manage-data/multitenancy/ | Multitenancy | Tiered Multitenancy > Configure Tiered Multitenancy | documentation | true | Multitenancy > Tiered Multitenancy > Configure Tiered Multitenancy
To take advantage of tiered multitenancy, you need to create a collection with user-defined sharding and create a fallback shard in it.
```python
client.create_collection(
collection_name="{collection_name}",
shard_number=1,
sharding_metho... | [
-0.0009032056550495327,
0.01954694278538227,
-0.03975916653871536,
0.0017626414773985744,
0.051066022366285324,
-0.05948759615421295,
0.054747872054576874,
0.059791449457407,
0.05372236296534538,
-0.016326695680618286,
0.043127402663230896,
0.030242841690778732,
0.04210088029503822,
-0.030... |
f2d6d1bd-e647-5ddf-aa19-bf7289bd3e37 | /documentation/tutorials-basics/generate-embeddings-fastembed/ | FastEmbed: Qdrant's Efficient Python Library for Embedding Generation | Next Steps | documentation | false | FastEmbed: Qdrant's Efficient Python Library for Embedding Generation > Next Steps
- **Pick another model.** FastEmbed also supports sparse (`SparseTextEmbedding`, including BM25, SPLADE, and miniCOIL), multi-vector (`LateInteractionTextEmbedding`, including ColBERT), and image (`ImageEmbedding`) embeddings, plus cros... | [
0.04540936276316643,
-0.03224354237318039,
-0.0428597629070282,
-0.020799875259399414,
0.07690956443548203,
-0.05685032531619072,
0.0692562535405159,
0.016476966440677643,
0.041408736258745193,
-0.026913078501820564,
0.04162754863500595,
-0.002229589968919754,
0.044052910059690475,
-0.0371... |
f2db3113-7479-52f8-a33b-21c938f6049d | /course/essentials/day-1/pitstop-project/ | Project: Building a Semantic Search Engine | course | false | Project: Building a Semantic Search Engine
Day 1
Now that you've seen how semantic search works with movies, it's time to build your own. Choose a domain you care about and create a search engine that understands meaning, not just keywords. | [
0.020857207477092743,
-0.06639748066663742,
-0.049907833337783813,
-0.00659281387925148,
0.04444831237196922,
-0.04160431772470474,
0.04146239534020424,
0.018040763214230537,
0.037983763962984085,
-0.017725862562656403,
0.036124251782894135,
0.034159377217292786,
0.0917172059416771,
-0.058... | |
f3005e43-27ba-5024-a3a0-9b26fbd98df2 | /documentation/search-tuning/when-a-reranker-is-worth-it/ | When Is a Reranker Worth It? | documentation | false | When Is a Reranker Worth It?
Before you tune a reranker, use the [pre-tuning checks](/documentation/search-tuning/before-tuning-a-qdrant-collection/) to verify index state and set a labeled baseline.
Your candidate list can already contain documents your ranking never shows. Score those candidates as if they were per... | [
0.010152543894946575,
-0.053450532257556915,
-0.04510093480348587,
-0.03149372339248657,
0.02505503222346306,
-0.03343183547258377,
0.05044661834836006,
0.023262565955519676,
0.0756431594491005,
-0.010039092972874641,
-0.015052255243062973,
0.02500854805111885,
0.036184243857860565,
-0.023... | |
f3145418-0896-54c2-a027-ef51e24463d4 | /articles/universal-io/ | Universal I/O in Qdrant | The Best of All Worlds (Almost) | article | false | Universal I/O in Qdrant > The Best of All Worlds (Almost)
Universal I/O started with a narrow question, which was how to read the data structures we already have from storage that isn't a local disk, and the answer turned out to be less about any single backend and more about finding the right interface between them.
... | [
0.03766680508852005,
-0.029831478372216225,
-0.05873384326696396,
0.000554094382096082,
0.07478013634681702,
-0.035361360758543015,
0.06076357886195183,
0.03263166546821594,
0.06064380332827568,
-0.020337937399744987,
0.023513376712799072,
0.02326168864965439,
0.06673308461904526,
0.007737... |
f31b24d9-fe0d-574a-b83a-dbc496a66d41 | /course/multi-vector-search/module-2/ | Module 2: Multi-Vector Representations for Multi-Modal Data | course | false | Module 2: Multi-Vector Representations for Multi-Modal Data
Module 2
Extend multi-vector representations beyond text to unlock powerful multi-modal search capabilities.
---
## Today's path
1. How ColPali Models Work
2. ColPali Family Overview
3. Visual Interpretability of ColPali
You'll learn to build multi-modal... | [
0.029176119714975357,
-0.03711014240980148,
-0.03475642949342728,
-0.03666192665696144,
0.07402196526527405,
-0.03438091278076172,
0.05315195024013519,
0.05089728534221649,
0.03172190859913826,
-0.002076722914353013,
0.039915021508932114,
0.03420577198266983,
0.06412048637866974,
-0.039568... | |
f3222e20-fe4f-5c07-8b9f-a88d0154274f | /documentation/examples/graphrag-qdrant-neo4j/ | GraphRAG with Qdrant and Neo4j | Implementation > Retrieval & Generation > End-to-End Pipeline | documentation | true | GraphRAG with Qdrant and Neo4j > Implementation > Retrieval & Generation > End-to-End Pipeline
Finally, let’s integrate everything into an end-to-end pipeline where we ingest some sample data, run the retrieval process, and query the language model.
```python
if __name__ == "__main__":
print("Script started")
... | [
0.02781420201063156,
-0.04077345132827759,
-0.05379227548837662,
-0.03769311681389809,
0.06974530965089798,
-0.03430984914302826,
0.05583224073052406,
0.06901238113641739,
0.06274820119142532,
-0.04166550934314728,
0.020367339253425598,
0.033506616950035095,
0.054112114012241364,
-0.030477... |
f32b6eb6-2a39-5fd5-8f3f-c765b7bb8665 | /documentation/migration-guidance/search-quality/ | Search Quality | Tier 2: Statistical Sampling (Recommended) > Why Recall Won't Be 1.0 (And That's OK) | documentation | false | Search Quality > Tier 2: Statistical Sampling (Recommended) > Why Recall Won't Be 1.0 (And That's OK)
Even a correct migration will often show recall@10 between 0.85 and 0.95 rather than 1.0. This isn't a bug. Here's why:
* **HNSW is approximate:** Both systems use approximate nearest neighbor algorithms. Different H... | [
0.045440394431352615,
-0.06187750771641731,
-0.05310901626944542,
-0.017075873911380768,
0.06842710077762604,
-0.014839744195342064,
0.07960469275712967,
0.03299493342638016,
0.10115062445402145,
-0.03525645285844803,
0.03578028082847595,
0.00968215987086296,
0.052409082651138306,
0.011525... |
f32be5b4-3f4f-57c3-a523-9a37cd028286 | /documentation/migrate-to-qdrant/from-pinecone/ | From Pinecone | Run the Migration | documentation | true | From Pinecone > Run the Migration
```bash
docker run --net=host --rm -it registry.cloud.qdrant.io/library/qdrant-migration pinecone \
--pinecone.index-host 'https://your-index-host.pinecone.io' \
--pinecone.index-name 'your-index' \
--pinecone.api-key 'pcsk_...' \
--qdrant.url 'https://your-instance.cl... | [
0.0645112544298172,
-0.06426019966602325,
-0.018822627142071724,
-0.019200986251235008,
0.056109458208084106,
-0.05365025997161865,
0.06336236000061035,
0.04848729819059372,
0.050814904272556305,
-0.04310561716556549,
0.03511224314570427,
-0.005487977992743254,
0.021254001185297966,
-0.000... |
f33c3bde-b0be-5e86-b9df-598e35fc12e6 | /documentation/tutorials-build-essentials/agentic-rag-camelai-discord/ | Discord RAG Bot | Step 5: Configure Discord | documentation | false | Discord RAG Bot > Step 5: Configure Discord
In the [Discord Developer Portal](https://discord.com/developers/applications), create an application and obtain its bot token from the **Bot** settings. Save that token as `DISCORD_BOT_TOKEN`.
Enable **Message Content Intent** in the portal. The message handler enables it ... | [
0.028732648119330406,
-0.024155709892511368,
-0.031240319833159447,
0.014521433040499687,
0.04246890917420387,
-0.05006527900695801,
0.036581579595804214,
0.05307407304644585,
0.026604821905493736,
-0.0066709984093904495,
0.03429824858903885,
-0.00828362163156271,
0.06058947741985321,
-0.0... |
f34e2d30-98a7-5bdd-bcc1-fb9fc12bbf0c | /documentation/tutorials-operations/large-scale-search/ | Large-Scale Search | Hardware | documentation | false | Large-Scale Search > Hardware
After some initial experiments, we figured out a minimal hardware configuration for the task:
- 8 CPU cores
- 64Gb RAM
- 1TB Disk space

This configuration is enough to index and explore the dataset in a... | [
-0.02263786457479,
-0.04542526602745056,
-0.05029884725809097,
-0.01697354204952717,
0.08197499811649323,
-0.046238742768764496,
0.019711744040250778,
0.04681212082505226,
0.07899350672960281,
-0.030699513852596283,
-0.0010527910199016333,
0.017949983477592468,
0.09128708392381668,
-0.0308... |
f35704e0-3854-5ea0-8b96-f47702fb3115 | /course/essentials/day-1/ | Day 1: Vector Search Fundamentals | course | false | Day 1: Vector Search Fundamentals
Day 1
Today we dive straight into Qdrant’s core data model and how vectors are compared.
---
## Today’s path
1. Points, Vectors and Payloads
2. Distance Metrics
3. Text Chunking Strategies
4. Demo: Semantic Movie Search
5. Project: Building a Semantic Search Engine
You’ll apply t... | [
0.05311010032892227,
-0.060315512120723724,
-0.036231059581041336,
0.008775767870247364,
0.047115880995988846,
-0.03628440946340561,
0.05920429900288582,
0.017630787566304207,
0.08959789574146271,
-0.04202510789036751,
0.04502008482813835,
0.02691284753382206,
0.048871785402297974,
-0.0439... | |
f376d22b-8326-530c-b6d3-b2172181996d | /documentation/migrate-to-qdrant/ | Migration Tool | General Advice | documentation | false | Migration Tool > General Advice
1. **Run the tool close to your databases.** Direct connectivity between source and target is not required — the tool streams through the machine it runs on. For best performance, use a machine with low latency to both.
2. **Use `--net=host` for local instances.** If either database ru... | [
0.016349848359823227,
-0.09395017474889755,
-0.037850189954042435,
-0.014006862416863441,
0.04090263321995735,
-0.02979382686316967,
0.050264615565538406,
0.011216195300221443,
0.05166121944785118,
-0.005176175385713577,
0.010237750597298145,
-0.02943352423608303,
0.04312234744429588,
-0.0... |
f3a91202-334f-5432-9ba6-5b287d4862c6 | /articles/distance-based-exploration/ | Data Exploration with Qdrant's Distance Matrix API | Dimensionality Reduction | article | true | Data Exploration with Qdrant's Distance Matrix API > Dimensionality Reduction
Initially, we might want to visualize an entire dataset, or at least a large portion of it, at a glance. However, high-dimensional data cannot be directly visualized. We must apply dimensionality reduction techniques to convert data into a l... | [
0.08407788723707199,
-0.026667486876249313,
-0.02395031787455082,
0.006661287508904934,
0.06889884918928146,
-0.030684005469083786,
0.08025109767913818,
0.050025954842567444,
0.08050327003002167,
-0.025165723636746407,
0.07483786344528198,
0.019085479900240898,
0.03709232807159424,
-0.0173... |
f3b39913-17b7-507b-baa3-f12ed30eb3de | /documentation/production-operations/memory-tiers/ | Memory Tiers in Qdrant: What to Use and When | documentation | false | Memory Tiers in Qdrant: What to Use and When
A growing vector collection eventually outgrows the RAM it started with: Qdrant handles that by letting you assign dense vectors, the HNSW graph, quantized vectors, payloads, and payload indexes each to whichever memory tier that structure supports, instead of forcing one R... | [
0.041305527091026306,
-0.05756007134914398,
-0.029058760032057762,
-0.046084772795438766,
0.04321996122598648,
-0.052578166127204895,
0.07002469897270203,
0.06601791828870773,
0.10298459976911545,
-0.027642108500003815,
0.018640916794538498,
0.04036322981119156,
0.03421658277511597,
0.0092... | |
f3c559d9-d038-55e3-bd83-0dc1dbbc855a | /documentation/migrate-to-qdrant/from-mongodb/ | From MongoDB | Run the Migration | documentation | true | From MongoDB > Run the Migration
```bash
docker run --net=host --rm -it registry.cloud.qdrant.io/library/qdrant-migration mongodb \
--mongodb.url 'mongodb://localhost:27017' \
--mongodb.database 'your-database' \
--mongodb.collection 'your-collection' \
--mongodb.vector-fields 'embedding' \
--qdran... | [
0.05965983495116234,
-0.05230500549077988,
-0.02158709615468979,
-0.03951818123459816,
0.09159141778945923,
-0.03199341893196106,
0.07136372476816177,
0.051518265157938004,
0.037222858518362045,
-0.05820579454302788,
0.04601587727665901,
0.0006544070201925933,
0.01748432219028473,
-0.01352... |
f3ccbe7c-657c-5399-88cb-7c751b9a09b2 | /documentation/cloud/cluster-monitoring/ | Monitor Clusters | Alerts | documentation | false | Monitor Clusters > Alerts
A single collection with payload index partitioning is usually optimal compared to many small individual tenant collections.*
It is also possible to split collections across clusters, the [Qdrant Migration CLI](/documentation/migrate-to-qdrant/) can help you with this.
**Where can I learn m... | [
0.01510187890380621,
-0.015229607932269573,
-0.047237757593393326,
-0.0033525486942380667,
0.044504497200250626,
-0.033653996884822845,
0.030845478177070618,
0.06606195867061615,
0.08979978412389755,
-0.059479739516973495,
0.014935433864593506,
0.04098540171980858,
0.015629829838871956,
-0... |
f3e2b30e-f2fb-5a05-8ce4-c8c535a0d28b | /documentation/embeddings/gemini/ | Gemini | Task Types | documentation | true | Gemini > Task Types
`gemini-embedding-2` does not accept the `task_type` parameter. Write the task into the text instead, and use the same format for every document and query in a collection:
| Use Case | Query Format | Document Format |
|---|---|---|
| Search | `task: search result \| query: {content}` | `title: {ti... | [
0.06372540444135666,
-0.029015304520726204,
-0.06365242600440979,
-0.030788762494921684,
0.04948660358786583,
-0.035278379917144775,
0.02250179648399353,
0.028182577341794968,
0.06130769103765488,
-0.012859052047133446,
0.03055073320865631,
0.00987967662513256,
0.043292295187711716,
-0.007... |
f3f2dbb8-8234-57ac-b965-13705d9f07da | /documentation/tutorials-develop/hybrid-search-fastembed/ | Build a Hybrid Search API | Workflow | documentation | false | Build a Hybrid Search API > Workflow
To create a hybrid search service, you will need to transform your raw data and then create a search function to manipulate it.
First, you will 1) download and prepare a sample dataset using a modified version of the BERT ML model. Then, you will 2) load the data into Qdrant, 3) c... | [
0.024991851300001144,
-0.02041907049715519,
-0.06419337540864944,
-0.019775159657001495,
0.07579842209815979,
-0.027817539870738983,
0.07197954505681992,
0.03474395349621773,
0.037418536841869354,
-0.00788997020572424,
0.04288671910762787,
0.006437564734369516,
0.0968000516295433,
-0.03075... |
f457cf71-3854-502f-ab9b-d9a4e6255b0f | /documentation/embeddings/voyage/ | Voyage AI | Models | documentation | false | Voyage AI > Models
| Model | Dimensions | Context | Input |
|---|---|---|---|
| `voyage-4-large` | 1024 (default), 2048, 512, 256 | 32,000 tokens | Text |
| `voyage-4` | 1024 (default), 2048, 512, 256 | 32,000 tokens | Text |
| `voyage-4-lite` | 1024 (default), 2048, 512, 256 | 32,000 tokens | Text |
| `voyage-code-4`... | [
0.02884627692401409,
-0.010998954065144062,
-0.0514923557639122,
-0.05555998161435127,
0.06413570046424866,
-0.0656399056315422,
0.06730883568525314,
0.04258529841899872,
0.055863428860902786,
-0.05318828672170639,
0.024866590276360512,
0.015779558569192886,
0.012527630664408207,
-0.011523... |
f47ec756-c617-5796-9afa-30bb9bfd6667 | /documentation/tutorials-operations/blue-green-deployment/ | Blue-Green Cluster Deployment | How It Works | documentation | false | Blue-Green Cluster Deployment > How It Works
The process has five stages:
1. Create a new cluster.
2. Load all data into it.
3. Test and validate the new cluster.
4. Switch application traffic to the new cluster.
5. (Optional) Decommission the old cluster. | [
-0.00908175390213728,
-0.05381238833069801,
-0.01899421587586403,
-0.048616472631692886,
0.09158245474100113,
-0.01941608637571335,
0.005473768338561058,
0.04336543753743172,
0.05581246688961983,
-0.015871891751885414,
0.00038214269443415105,
0.030099891126155853,
-0.007127491291612387,
-0... |
f4857d98-b37c-521e-8e88-9bb91db4b138 | /course/beginners/module-2/core-data-model/ | Core Data Model | course | false | Core Data Model
Module 2
Qdrant organizes data in three levels. Understanding this structure is the foundation for everything else in the course.
 | [
0.040603749454021454,
-0.013584369793534279,
-0.032629162073135376,
-0.049597062170505524,
0.06508661061525345,
-0.038575250655412674,
0.06242183968424797,
0.10553120821714401,
0.10206802189350128,
-0.0058727748692035675,
0.020309319719672203,
0.007009726017713547,
0.08032465726137161,
-0.... | |
f48a8697-e7d9-5bc4-b9ea-c674e0499710 | /articles/relevance-feedback/ | Relevance Feedback in Qdrant | Feedback-Based Scoring | article | false | Relevance Feedback in Qdrant > Feedback-Based Scoring
With the context pair(s) at our expense, we can try the following feedback-based scoring during retrieval:
1. Let the retriever still have a say in what is relevant to the query, count in its **score** (to query).
2. Yet reward candidates that are closer to the ... | [
0.02375698648393154,
-0.051682520657777786,
-0.024127816781401634,
-0.052089206874370575,
0.028668638318777084,
-0.033121638000011444,
0.04814239218831062,
0.0074021583423018456,
0.06510786712169647,
0.002464024815708399,
0.024504603818058968,
0.036294370889663696,
0.07449574023485184,
-0.... |
f48d6e61-a77a-52fb-b735-c730e2bd2be4 | /articles/filtered-vector-search-acorn/ | Filtered Vector Search: What ACORN Fixes, and What Fixes ACORN | The Two ACORNs | article | false | Filtered Vector Search: What ACORN Fixes, and What Fixes ACORN > The Two ACORNs
The [ACORN paper](https://arxiv.org/abs/2403.04871) (Patel et al., SIGMOD 2024) describes two algorithms. Its headline claim of "2-1,000x higher throughput at a fixed recall" belongs to ACORN-gamma, which expands neighbor lists during inde... | [
0.051289476454257965,
-0.07714386284351349,
-0.06068594008684158,
-0.03499365225434303,
0.06773751229047775,
0.02082352340221405,
0.059284355491399765,
0.035073135048151016,
0.05460219085216522,
0.007877564057707787,
0.05658382922410965,
0.032207295298576355,
0.0319233164191246,
-0.0149102... |
f4b8369c-3521-5fe7-b7d1-731af7d1b02b | /documentation/cloud/configure-cluster/ | Configure Clusters | Advanced Optimizations | documentation | false | Configure Clusters > Advanced Optimizations
Configuring these advanced settings will have an impact on performance and reliability. We recommend using the default values unless you are confident they are required for your use case.
*Optimizer CPU Budget*
Configures how many CPUs (threads) to allocate for optimizatio... | [
0.025079697370529175,
-0.04082557186484337,
-0.029064210131764412,
0.006486371159553528,
0.06059100851416588,
-0.04207379370927811,
0.010613302700221539,
0.028838932514190674,
0.06155228242278099,
-0.033101219683885574,
0.026676148176193237,
-0.021774480119347572,
0.04340408742427826,
-0.0... |
f4c92d61-09b5-5830-8347-18f4e9f432a5 | /articles/sparse-vectors/ | Understanding SPLADE and Sparse Vectors | Conclusion | article | false | Understanding SPLADE and Sparse Vectors > Conclusion
Alright, folks, let's wrap it up. Better search isn't a 'nice-to-have,' it's a game-changer, and Qdrant can get you there.
Got questions? Our [Discord community](https://qdrant.to/discord?utm_source=qdrant&utm_medium=website&utm_campaign=sparse-vectors&utm_content=... | [
0.022561397403478622,
-0.033885400742292404,
-0.05210326984524727,
-0.027108240872621536,
0.025403713807463646,
-0.039451081305742264,
0.06364700943231583,
0.034345876425504684,
0.06692337244749069,
-0.034951742738485336,
0.016177508980035782,
0.03803853318095207,
0.040920112282037735,
0.0... |
f4caf9fd-48bd-588e-87d6-dbc79cb3333b | /documentation/tutorials-search-engineering/using-relevance-feedback/ | Relevance Feedback Retrieval in Qdrant | Customizing for Your Use Case > Setup > Dataset | documentation | true | Relevance Feedback Retrieval in Qdrant > Customizing for Your Use Case > Setup > Dataset
We'll use one of the datasets available in the Qdrant Cluster UI.
1. Go to Cluster UI -> Datasets.
2. Out of 3 datasets there, pick the first one, "Qdrant Web Documentation". It consists of small text chunks from [our documentat... | [
0.05175461620092392,
-0.058230895549058914,
-0.026129651814699173,
-0.03534184768795967,
0.053277332335710526,
-0.031457509845495224,
0.03919544816017151,
0.0321955643594265,
0.03750988095998764,
-0.04944887012243271,
0.006925542838871479,
0.0319061279296875,
0.05500153824687004,
-0.053269... |
f4cee759-7413-5275-a314-ba28fec025d8 | /course/multi-vector-search/module-2/visual-interpretability/ | Visual Interpretability of ColPali | Practical Example: Debugging a Search | course | true | Visual Interpretability of ColPali > Practical Example: Debugging a Search
Visual interpretability becomes powerful when debugging search results. Let's walk through a complete example to understand why certain documents match (or don't match) specific queries.
### Scenario: Investigating an Unexpected Match
Imagine... | [
0.0112838139757514,
-0.0548643134534359,
-0.0670868456363678,
-0.04338851198554039,
0.07496034353971481,
-0.001539462013170123,
0.03736317530274391,
0.028382686898112297,
0.025470221415162086,
-0.03490114212036133,
0.022772805765271187,
0.013116436079144478,
0.07087378203868866,
-0.0203683... |
f4d0531c-10e3-5251-9bae-af548d642477 | /documentation/fastembed/fastembed-postprocessing/ | Multi-Vector Postprocessing | Using MUVERA in FastEmbed > Load Model and Create MUVERA Processor | documentation | true | Multi-Vector Postprocessing > Using MUVERA in FastEmbed > Load Model and Create MUVERA Processor
Load the ColBERT model and wrap it with a MUVERA processor.
```python
model_name = "colbert-ir/colbertv2.0"
model = LateInteractionTextEmbedding(model_name=model_name)
# Create MUVERA processor with recommended parameter... | [
0.028438862413167953,
-0.03874380886554718,
-0.015455329790711403,
-0.06036781519651413,
0.0677480399608612,
-0.04373593628406525,
0.0787239521741867,
0.04737077280879021,
0.051147788763046265,
0.003952504135668278,
0.024757197126746178,
0.02051614597439766,
0.029451537877321243,
-0.028439... |
f4f4c6a6-451e-551f-befd-a713d83fcb98 | /documentation/manage-data/indexing/ | Indexing | Full-Text Index > Tokenizers > Code | documentation | false | Indexing > Full-Text Index > Tokenizers > Code
With [phrase matching](#phrase-matching), a joined token only matches a phrase that consists of that token alone. For example, the phrase `hashmap new` doesn't match `HashMap::new`, but `hash map new` does. | [
0.023846430703997612,
-0.026658883318305016,
-0.07154328376054764,
-0.07068764418363571,
0.054782163351774216,
-0.0576678141951561,
0.0352594368159771,
0.00305860280059278,
0.047471705824136734,
-0.005987491458654404,
0.047474298626184464,
0.02393745444715023,
0.03660990670323372,
-0.00088... |
f4f681f8-d99d-554c-a11a-bfb69bc01b69 | /course/essentials/day-6/final-project/ | Final Project: Production-Ready Documentation Search Engine | Success Criteria | course | false | Final Project: Production-Ready Documentation Search Engine > Success Criteria
Your project succeeds when it shows production-ready search with measurable performance.
Complete end-to-end Jupyter notebook or app that runs against Qdrant Cloud
Hybrid search implementation with dense, sparse, and multivector compone... | [
0.03347790613770485,
-0.03734331950545311,
-0.032989829778671265,
-0.052272677421569824,
0.06409472227096558,
-0.04179512709379196,
0.10239338129758835,
0.016433609649538994,
0.08412998169660568,
-0.042086414992809296,
0.006824385840445757,
0.08069535344839096,
0.07958637923002243,
-0.0305... |
f4fbb1bc-9478-5792-8568-bcff36312183 | /course/essentials/day-7/unstructured/ | Integrating with Unstructured.io | Resources | course | false | Integrating with Unstructured.io > Resources
- [Unstructured Qdrant Destination](https://docs.unstructured.io/ui/destinations/qdrant):
Official Unstructured documentation for sending processed data to Qdrant. Learn about Qdrant Cloud integration, collection setup, and workflow configuration.
- [Qdrant & Unstructu... | [
0.01556667871773243,
-0.028080830350518227,
-0.07653921097517014,
-0.028946230188012123,
0.043799083679914474,
-0.029033437371253967,
0.046888817101716995,
0.04826171323657036,
0.031385116279125214,
-0.019246665760874748,
-0.016459936276078224,
-0.012004131451249123,
0.04022477567195892,
-... |
f52feaed-adf2-5f32-ae02-4cf0aa6b8e9c | /articles/relevance-feedback/ | Relevance Feedback in Qdrant | Vector Search Optimization > Tools of a Vector Search Engine | article | false | Relevance Feedback in Qdrant > Vector Search Optimization > Tools of a Vector Search Engine
The goal is to fill this gap and let search end-users (doesn't matter if humans or agents) get a proper relevance feedback interface to use in our vector search engine.
So what makes a method a production-ready for vector sear... | [
-0.0000020326215235400014,
-0.05911385640501976,
-0.06969930231571198,
-0.05470496043562889,
0.06291064620018005,
-0.011240148916840553,
0.0808214396238327,
0.020513959228992462,
0.07287942618131638,
0.0022927895188331604,
0.025306284427642822,
0.03350913152098656,
0.092837393283844,
-0.02... |
f5448100-703b-57e4-94f3-571363ca471f | /course/essentials/ | Qdrant Essentials Course | What you'll learn | course | false | Qdrant Essentials Course > What you'll learn
- Qdrant data modeling: points, payloads, and schemas
- Embeddings, chunking, and similarity metrics
- Indexing and retrieval tuning (HNSW, filters, recall/latency)
- Hybrid search with sparse + dense vectors and re-ranking
- Performance optimization, compression, and quant... | [
0.059482160955667496,
-0.019805677235126495,
-0.0555400550365448,
-0.04406851530075073,
0.06678469479084015,
-0.05393077805638313,
0.045260921120643616,
0.03612757846713066,
0.09408780187368393,
-0.04435931146144867,
0.00397526053711772,
-0.008356167934834957,
0.03920811414718628,
-0.03860... |
f5609d09-e786-5027-a753-8deffe635576 | /articles/hyperbolic-embeddings-qdrant/ | Why Hyperbolic Embeddings Matter for Hierarchical Data | Serving Them with Qdrant > What About a Production 512-Dimensional Model? | article | false | Why Hyperbolic Embeddings Matter for Hierarchical Data > Serving Them with Qdrant > What About a Production 512-Dimensional Model?
One property makes these models easier to serve than expected. For a model on the hyperboloid, ranking by the Lorentz inner product is rank identical to ranking by geodesic distance. Acros... | [
0.061505384743213654,
0.008684989996254444,
-0.03238062188029289,
-0.01764507405459881,
0.07935034483671188,
-0.021306799724698067,
0.062152132391929626,
0.03283082693815231,
0.10063730925321579,
-0.022637058049440384,
0.010061615146696568,
0.02163321152329445,
0.05375290289521217,
-0.0279... |
f5c8e012-debf-5285-a2b6-b5f15e91e468 | /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 |
|-----------------------------------... | [
0.006910355761647224,
-0.025773126631975174,
-0.029648110270500183,
0.00840621255338192,
0.06822226941585541,
-0.06620558351278305,
0.06863731890916824,
0.05911914259195328,
0.05941998213529587,
-0.02726863883435726,
0.03589383140206337,
-0.005141805857419968,
0.014008982107043266,
-0.0347... |
f5ce1c31-e38c-5ab0-a994-f8ef9086fe1c | /documentation/tutorials-basics/reranking-hybrid-search/ | Hybrid Search with Reranking | Implementation > Create Collection | documentation | true | Hybrid Search with Reranking > Implementation > Create Collection
Create a new collection called `hybrid-search`, configured with two vector types and a full-text index:
- **Dense embeddings** (`dense`) using cosine distance for semantic comparisons.
- **Late interaction embeddings** (`multi`) using cosine distance, ... | [
0.06353253871202469,
-0.03142940253019333,
-0.06913194060325623,
-0.05180062726140022,
0.09024466574192047,
-0.03523354232311249,
0.046974457800388336,
0.04524872824549675,
0.07802939414978027,
-0.029242148622870445,
0.052331164479255676,
0.04669144004583359,
0.05665808171033859,
-0.011575... |
f5e70d7b-a2a3-51be-aa6d-595cee8d6079 | /articles/metric-learning-tips/ | Metric Learning Tips & Tricks | Conclusion | article | false | Metric Learning Tips & Tricks > Conclusion
In this article, I have shown how metric learning can be more scalable and flexible than the classification models.
I suggest trying similar approaches in your tasks - it might be matching similar texts, images, or audio data.
With the existing variety of pre-trained neural n... | [
0.006221221759915352,
0.008531679399311543,
-0.043292608112096786,
-0.01707298308610916,
0.07031422108411789,
-0.04707440361380577,
0.06898816674947739,
0.0636199563741684,
0.060176219791173935,
0.002318492392078042,
0.02484002336859703,
0.029844794422388077,
0.09808304905891418,
-0.010986... |
f6015c9e-5207-5ad9-82e9-e0ee5fc4fdef | /documentation/tutorials-build-essentials/agentic-rag-crewai-zoom/ | Agentic RAG with CrewAI | What You'll Build | documentation | false | Agentic RAG with CrewAI > What You'll Build
In this hands-on tutorial, we'll create a system that:
1. Uses Qdrant to store and retrieve meeting transcripts as vector embeddings
2. Leverages CrewAI agents to analyze and summarize meeting data
3. Presents insights in a simple Streamlit interface for easy interaction
... | [
0.058826129883527756,
-0.05426684021949768,
-0.045193497091531754,
-0.03470601886510849,
0.0756671205163002,
-0.056031446903944016,
0.06397225707769394,
0.0513627752661705,
0.04207479953765869,
-0.036899909377098083,
0.016976388171315193,
-0.00545011879876256,
0.0493883453309536,
-0.028226... |
f62041c1-63f4-5dd3-b5a1-125fb463fbb3 | /documentation/embeddings/nvidia/ | NVIDIA | Embed Documents | documentation | true | NVIDIA > Embed Documents
Embed documents with `input_type` set to `passage`. Each item in the response carries the `index` of its input, which this example uses as the point ID:
```python
documents = [
"Qdrant is a vector search engine.",
"NVIDIA Nemotron embedding models turn text into dense vectors.",
]
re... | [
0.03691282868385315,
-0.05886230990290642,
-0.0616757906973362,
-0.05949241667985916,
0.0704386979341507,
-0.05518375709652901,
0.04904331639409065,
0.08074100315570831,
0.08022788912057877,
-0.03565310686826706,
0.00826067104935646,
0.04083770513534546,
-0.005833682604134083,
-0.017989085... |
f630076c-07f8-5db9-bb6d-4771a85ed8a5 | /documentation/data-management/poma/ | POMA | 4. Chunk a File with POMA | documentation | true | POMA > 4. Chunk a File with POMA
```python
client = Poma(os.environ["POMA_API_KEY"])
job = client.start_chunk_file("./docs/your_file.pdf")
chunk_data = client.get_chunk_result(
job["job_id"],
show_progress=True,
download_dir="./",
filename="your_file.poma",
)
```
### POMA-specific knobs in get_chunk_... | [
0.020054958760738373,
-0.06815750896930695,
-0.020586302503943443,
-0.08016538619995117,
0.06441856175661087,
-0.02378253825008869,
0.05285836383700371,
0.017546135932207108,
0.06772410124540329,
-0.010194843634963036,
0.01964413747191429,
0.015540710650384426,
0.058363065123558044,
-0.054... |
f6376fe2-f51f-59c9-bc05-f01f3cb8a02c | /documentation/platforms/ | Platforms | documentation | false | Platforms
## Platform Integrations
| Platform | Description |
| ------------------------------------------------------- | ------------------------------------------------------------------------... | [
0.0070772659964859486,
-0.054574377834796906,
-0.053237561136484146,
-0.02369232103228569,
0.07357211410999298,
-0.027334334328770638,
0.04723938927054405,
0.017209332436323166,
0.04139728099107742,
-0.028831282630562782,
0.011646948754787445,
0.02606678567826748,
0.0012486919295042753,
-0... | |
f6429c10-bea8-55f9-a3f8-b59ba6594ee3 | /documentation/fastembed/fastembed-colbert/ | Working with ColBERT | Upload embeddings to Qdrant | documentation | true | Working with ColBERT > Upload embeddings to Qdrant
Install `qdrant-client`
```python
pip install "qdrant-client>=1.14.2"
```
Qdrant Client has a simple in-memory mode that allows you to experiment locally on small data volumes.
Alternatively, you could use for experiments [a free cluster](https://qdrant.tech/docume... | [
0.045856721699237823,
-0.037083789706230164,
-0.05505625903606415,
-0.03699932247400284,
0.09425900876522064,
-0.03563609719276428,
0.07908068597316742,
0.03340131416916847,
0.07449349015951157,
-0.02206822670996189,
0.042202822864055634,
-0.004000415559858084,
0.07668443769216537,
-0.0407... |
f64470b4-e0f7-50d8-a210-cae1ee7a0faa | /documentation/search/hybrid-queries/ | Hybrid Queries | Multi-Stage Queries > Fusion in Distributed Collections | documentation | false | Hybrid Queries > Multi-Stage Queries > Fusion in Distributed Collections
The previous example puts the fusion inside a prefetch. In a multi-shard collection, a fusion merges results across all shards only when it is the main query, held in the top-level `query` field with the retrievers as its prefetches, as in the [R... | [
0.022243520244956017,
-0.03389374539256096,
-0.029215658083558083,
-0.031140347942709923,
0.07213640958070755,
-0.04356842115521431,
0.06032061576843262,
0.03985922783613205,
0.049898531287908554,
-0.031346648931503296,
0.03175533190369606,
0.026812026277184486,
0.02662048675119877,
0.0044... |
f6469377-33b4-5294-9a48-a364a295e163 | /documentation/ops-monitoring/hybrid-cloud-prometheus/ | Monitoring with Grafana and Prometheus | Step 4: Access Grafana | documentation | true | Monitoring with Grafana and Prometheus > Step 4: Access Grafana
Once Prometheus is configured to scrape metrics from Qdrant, you can access Grafana to visualize the metrics.
Get the Grafana 'admin' user password by running:
```bash
kubectl --namespace monitoring get secrets prometheus-grafana -o jsonpath="{.data.adm... | [
0.024203691631555557,
-0.01735440082848072,
-0.025122705847024918,
0.018034856766462326,
0.05348711460828781,
-0.04012170806527138,
0.057811688631772995,
0.07715653628110886,
0.016010582447052002,
-0.004363432992249727,
0.05439525097608566,
-0.0067921895533800125,
0.08884453773498535,
-0.0... |
f64d6b67-8e0e-56f3-9c34-3e89032075b2 | /documentation/scaling/consistency-guarantees/ | Consistency Guarantees | Read Consistency | documentation | true | Consistency Guarantees > Read Consistency
Read `consistency` can be specified for most read requests and will ensure that the returned result
is consistent across cluster nodes.
- `all` will query all nodes and return points, which present on all of them
- `majority` will query all nodes and return points, which pres... | [
0.04614179953932762,
-0.047883290797472,
-0.053771376609802246,
-0.0403696708381176,
0.05104641243815422,
-0.03730863705277443,
0.07702121138572693,
0.0417981781065464,
0.04204206168651581,
-0.00008475073263980448,
0.049910131841897964,
0.05107647180557251,
0.044720474630594254,
0.00571461... |
f65c14f0-7ad7-5ccd-bebc-60b946302882 | /documentation/edge/edge-quickstart/ | Quickstart | Filter points | documentation | true | Quickstart > Filter points
You can also [filter](/documentation/search/filtering/) points based on payload fields:
```python
from qdrant_edge import FieldCondition, Filter, MatchValue
results = edge_shard.query(
QueryRequest(
query=Query.Nearest([0.2, 0.1, 0.9, 0.7], using=VECTOR_NAME),
filter=Fi... | [
0.029744217172265053,
-0.0018705109832808375,
-0.0028400160372257233,
-0.046872757375240326,
0.07350572198629379,
-0.026369431987404823,
0.024151571094989777,
0.056023191660642624,
0.08472141623497009,
-0.02220781520009041,
0.07528126239776611,
0.03482397645711899,
0.07558140903711319,
-0.... |
f65d935d-7db8-592b-a654-8c2fbbc74fac | /documentation/manage-data/quantization/ | Quantization | Quantization Tips > Accuracy Tuning | documentation | true | Quantization > Quantization Tips > Accuracy Tuning
In this section, we will discuss how to tune the search precision.
The fastest way to understand the impact of quantization on the search quality is to compare the search results with and without quantization.
In order to disable quantization, you can set `ignore` to... | [
0.03936098888516426,
-0.06412860751152039,
-0.04247350990772247,
-0.008248825557529926,
0.04749053344130516,
-0.002035583835095167,
0.033425331115722656,
0.022301824763417244,
0.08898567408323288,
-0.020378319546580315,
0.04344276338815689,
0.02873295731842518,
0.07398931682109833,
-0.0220... |
f664cb8d-516d-547d-bb03-f7887f758a1e | /documentation/search/text-search/sparse-vector-search/ | Sparse Vector Search | BM25 > BM25 Text Processing > ASCII Folding | documentation | true | Sparse Vector Search > BM25 > BM25 Text Processing > ASCII Folding
ASCII folding is the process of removing diacritics (accents) from characters. By removing diacritics, ASCII folding enables you to ignore accents when searching. For instance, with ASCII folding enabled, searching for "cafe" matches both "cafe" and "c... | [
0.051371216773986816,
-0.06342923641204834,
-0.020552832633256912,
-0.065633125603199,
0.04007629305124283,
-0.037646226584911346,
0.02565407194197178,
0.04536661133170128,
0.053827833384275436,
0.00036599638406187296,
0.024503810331225395,
-0.0030683584045618773,
0.05439446121454239,
0.00... |
f6970021-47f0-5b3b-9ba2-1678314098ec | /documentation/search/text-search/text-filtering/ | Text Filtering | Text and Keyword Strings | documentation | false | Text Filtering > Text and Keyword Strings
When it comes to filtering on strings, it is important to understand the difference between the two types of strings in Qdrant: text and keyword. These two string types are designed for different use cases: filtering on exact string values or filtering on individual search ter... | [
0.038891132920980453,
-0.039257679134607315,
-0.04142618924379349,
-0.032691940665245056,
0.05473941192030907,
-0.01968894898891449,
0.060238417237997055,
0.024869384244084358,
0.08996975421905518,
-0.030475758016109467,
0.0329262800514698,
0.003978564869612455,
0.05034308508038521,
-0.007... |
f6ad0e8a-7c91-55f8-a9ca-8715323abf2e | /course/multi-vector-search/module-2/visual-interpretability/ | Visual Interpretability of ColPali | Manual Implementation with FastEmbed | course | true | Visual Interpretability of ColPali > Manual Implementation with FastEmbed
Let's build a complete interpretability system from scratch using only FastEmbed and standard Python libraries. This approach works with any late interaction model and gives you full control over the visualization process.
### Step 1: Generate ... | [
0.03837643191218376,
-0.05207112431526184,
-0.04766499251127243,
-0.03334294632077217,
0.0823846161365509,
-0.05366264283657074,
0.07785961776971817,
0.05813358351588249,
-0.012659438885748386,
-0.005832267925143242,
0.057083334773778915,
0.011335412971675396,
0.05384153127670288,
-0.02841... |
f6bfa6e6-2df4-52fe-bc59-424e349d63c8 | /documentation/private-cloud/changelog/ | Changelog | 1.3.0 (2025-01-17) | documentation | false | Changelog > 1.3.0 (2025-01-17)
| Component | Version |
|------------------------|---------|
| qdrant-kubernetes-api | v1.8.0 |
| operator | 2.1.21 |
| qdrant-cluster-manager | v0.2.10 |
* Support for re-sharding with Qdrant >= 1.13.0 | [
0.02848942019045353,
-0.034218110144138336,
-0.06462126225233078,
-0.036401793360710144,
0.03920508176088333,
-0.048734668642282486,
0.05478142201900482,
0.049280453473329544,
0.04190325736999512,
-0.04422667995095253,
0.037131793797016144,
0.021583937108516693,
0.04561265930533409,
-0.047... |
f6c1c1fc-7f38-5639-8340-2a20cb98dbe9 | /course/multi-vector-search/module-3/quantization-techniques/ | Vector Quantization Techniques | Quantization: Compressing Without Losing Quality | course | false | Vector Quantization Techniques > Quantization: Compressing Without Losing Quality
**Vector quantization** reduces memory by representing vectors with fewer bits while preserving the relative distances between them. Qdrant supports several quantization methods optimized for different scenarios.
**Important: What Quant... | [
0.04594392329454422,
-0.0561770498752594,
-0.052606482058763504,
-0.05941535532474518,
0.046699315309524536,
-0.019783318042755127,
0.04137418791651726,
0.009065412916243076,
0.09721564501523972,
-0.036149006336927414,
0.03147032856941223,
0.03127991035580635,
0.03820132836699486,
-0.00702... |
f6ef9550-db74-5c42-8cef-149f6f791443 | /documentation/cloud-rbac/permission-reference/ | Permission Reference | Cluster > Clusters | documentation | false | Permission Reference > Cluster > Clusters
| Permission | Description |
|------------|------------|
| `read:clusters` | View cluster details. |
| `write:clusters` | Modify cluster settings. |
| `delete:clusters` | Delete clusters. | | [
0.023419713601469994,
-0.0062075029127299786,
-0.06931796669960022,
-0.061157405376434326,
0.06380301713943481,
-0.021078508347272873,
0.01200935710221529,
0.0729411318898201,
0.0637112557888031,
-0.03381313383579254,
0.041900385171175,
0.008286441676318645,
0.06823328137397766,
-0.0108832... |
f6f2687d-82ed-573e-bf55-ae62277239d8 | /course/essentials/day-3/sparse-retrieval-demo/ | Demo: Keyword Search with Sparse Vectors | Which Model to Choose | course | false | Demo: Keyword Search with Sparse Vectors > Which Model to Choose
| | **BM25** | **miniCOIL** | **SPLADE** |
|---|---|---|---|
| **Matching** | Exact term overlap only | Exact term overlap only | Exact terms + synonym-like expansion |
| **Domain support** | High, statistical model, works out of the box on any domain |... | [
0.03485749289393425,
-0.04319995641708374,
-0.06557348370552063,
-0.042924586683511734,
0.049320973455905914,
-0.052839167416095734,
0.07977021485567093,
0.018157489597797394,
0.06923722475767136,
-0.03852012753486633,
0.033130135387182236,
0.028045348823070526,
0.042319074273109436,
-0.01... |
f6ff49ca-76d6-5a9e-b7ec-ff3902e6fd49 | /documentation/tutorials-operations/blue-green-deployment/ | Blue-Green Cluster Deployment | Prerequisites | documentation | false | Blue-Green Cluster Deployment > Prerequisites
- A running Qdrant cluster serving production traffic (the blue cluster).
- Access to [Qdrant Cloud](https://cloud.qdrant.io/) or your self-hosted infrastructure to spin up a new cluster.
- If you use a private connection on Qdrant Cloud, read [Private Connections on Q... | [
-0.0002924124419223517,
-0.050893060863018036,
-0.04282834008336067,
-0.07064321637153625,
0.0731244683265686,
-0.021143538877367973,
0.05773671343922615,
0.03490125387907028,
0.04051407426595688,
-0.030781913548707962,
0.001590857980772853,
-0.005372443702071905,
0.018682172521948814,
-0.... |
f70be64f-cb68-5885-b59f-7fbf24502716 | /documentation/data-synchronization/with-postgres/ | Keeping Postgres and Qdrant in Sync | Tier 2: Transactional Outbox Pattern > The Worker | documentation | true | Keeping Postgres and Qdrant in Sync > Tier 2: Transactional Outbox Pattern > The Worker
A background async task processes outbox events in batches. The critical concurrency primitive is `FOR UPDATE SKIP LOCKED` — it lets multiple workers claim events safely without stepping on each other:
```py
async def process_batc... | [
0.01942945271730423,
-0.023890523239970207,
-0.0383460707962513,
-0.032179832458496094,
0.04314759373664856,
-0.05154243856668472,
0.07346110790967941,
-0.002686093794181943,
0.0755230262875557,
-0.008426085114479065,
0.03785371035337448,
0.06289029121398926,
0.044515930116176605,
-0.03526... |
f70c5de7-d01c-5275-b5c5-48e939e4b3e0 | /documentation/manage-data/storage/ | Storage | documentation | false | Storage
All data within one collection is divided into segments.
Each segment has its independent vector and payload storage as well as indexes.
Data stored in segments usually do not overlap.
However, storing the same point in different segments will not cause problems since the search contains a deduplication mecha... | [
0.0335216261446476,
-0.04512086883187294,
-0.03783972188830376,
-0.020178165286779404,
0.034603387117385864,
-0.007825581356883049,
0.052805423736572266,
0.07582930475473404,
0.0623125359416008,
0.00007882945646997541,
0.008288178592920303,
0.03285294398665428,
0.021681269630789757,
-0.002... | |
f73e33c4-b9d2-5bc2-be6c-722d1a41ff99 | /documentation/search-evaluation/pipeline-output-quality/ | Evaluating Pipeline Output Quality | Scoring with Ragas | documentation | true | Evaluating Pipeline Output Quality > Scoring with Ragas
Three Ragas metrics cover the common failure modes for pipeline output quality:
- **`faithfulness`** checks whether the answer only makes claims supported by the retrieved context. It drops when the generator hallucinates or uses its training knowledge instead o... | [
0.03612741827964783,
-0.01804768666625023,
-0.05264005810022354,
-0.0577235221862793,
0.03960325941443443,
-0.02956370823085308,
0.06404697895050049,
0.01583249866962433,
0.05775643140077591,
-0.05863203480839729,
0.007994520477950573,
0.009139695204794407,
0.06876779347658157,
-0.00836754... |
f7537022-d615-5e6b-a359-437fd4ac99b9 | /documentation/agentic-tools/skills/ | Agent Skills | Philosophy | documentation | false | Agent Skills > Philosophy
Skills are not a second copy of the documentation.
They act as a navigation and decision layer, helping an agent determine which docs apply, when a feature is the right choice, and what pitfalls to avoid.
Documentation answers **“how?”** Skills answer **“when?”**, **“why?”**, and **“what sho... | [
0.033501382917165756,
-0.04197593778371811,
-0.10519108921289444,
-0.05414775013923645,
0.03702160716056824,
-0.05128668621182442,
0.0247028935700655,
0.005612467881292105,
0.07837393879890442,
-0.021405916661024094,
0.032989658415317535,
0.0180796068161726,
0.05791604518890381,
-0.0258193... |
f7565cfc-f964-58b9-9bab-336ce554addc | /documentation/frameworks/smolagents/ | SmolAgents | documentation | true | SmolAgents
HuggingFace [SmolAgents](https://github.com/huggingface/smolagents) is a Python library for building AI agents. These agents write Python code to call tools and orchestrate other agents.
It uses `CodeAgent`. An LLM engine that writes its actions in code. SmolAgents suggests that this approach is demonstrat... | [
0.0393623486161232,
-0.024845311418175697,
-0.0657910406589508,
-0.029073093086481094,
0.0419728048145771,
-0.05243036895990372,
0.06004320830106735,
0.03734980896115303,
0.05402829870581627,
-0.034079957753419876,
0.058424901217222214,
0.01629732735455036,
0.026605088263750076,
-0.0381860... | |
f7675a70-baf9-582d-83b5-3a41652f62e2 | /documentation/overview/ | Overview | Scaling Considerations > Memory Requirements | documentation | false | Overview > Scaling Considerations > Memory Requirements
Memory is a critical resource when scaling vector search. By default, Qdrant stores vectors in RAM for maximum search performance, but as collections grow to millions of vectors, keeping everything in memory becomes expensive. Qdrant lets you control the memory u... | [
0.028066104277968407,
-0.054569046944379807,
-0.02769944816827774,
-0.03501208499073982,
0.04450523108243942,
-0.03535087779164314,
0.06790725886821747,
0.030799470841884613,
0.11949070543050766,
-0.02150697633624077,
-0.001733617507852614,
0.004069080576300621,
0.05397291108965874,
-0.013... |
f797005e-9b5f-5850-b6a2-520c467b7eb1 | /documentation/scaling/resilience/ | Resilience | documentation | false | Resilience
Qdrant's fault tolerance is driven by replication and node count. Together, they control whether your data survives a node loss and whether your cluster keeps serving requests when one goes down.
Resilience covers three distinct aspects of a Qdrant cluster, and it's worth keeping them separate since a clus... | [
0.021620478481054306,
-0.014003734104335308,
-0.0727534070611,
-0.04980171471834183,
0.06540553271770477,
-0.0313340462744236,
0.06343735009431839,
0.021718842908740044,
0.0634923130273819,
-0.022350141778588295,
0.05527065321803093,
0.003584594000130892,
0.06354399025440216,
-0.0146364001... | |
f7df213c-4347-529b-b38b-d26ea2ba6c27 | /documentation/tutorials-operations/embedding-model-migration/ | Migrate to a New Embedding Model | Blue-Green Migration > Step 2: Enable Dual Writes | documentation | true | Migrate to a New Embedding Model > Blue-Green Migration > Step 2: Enable Dual Writes
To ensure that both collections are kept up-to-date during the migration, write any changes to both collections simultaneously. This way, any new data or updates to existing data are reflected in both collections.
Ideally, the data i... | [
0.04205000773072243,
-0.048785485327243805,
-0.05085699260234833,
0.004747339524328709,
0.050248079001903534,
-0.03290335834026337,
0.05917752906680107,
0.0648558959364891,
0.038910917937755585,
-0.006888442672789097,
0.01748890057206154,
0.016303082928061485,
0.03138229623436928,
-0.02093... |
f7fd97aa-5df5-579f-9205-9bf0b1634b36 | /articles/dedicated-vector-search/ | Why Vector Search Needs a Dedicated Database | Objections to a Dedicated Service > "Every database will add vector search" | article | false | Why Vector Search Needs a Dedicated Database > Objections to a Dedicated Service > "Every database will add vector search"
Adding a vector index to a database does not change the architecture around it. A relational database is built for transactions and strong consistency, and its storage, replication, and scaling mo... | [
-0.017014821991324425,
-0.051394373178482056,
-0.053046826273202896,
-0.037282008677721024,
0.010895414277911186,
-0.03713352233171463,
0.05426056310534477,
0.04587412625551224,
0.06944829225540161,
-0.02098120003938675,
0.027971012517809868,
0.05910075828433037,
0.043468356132507324,
-0.0... |
f819d322-0c9f-5670-936b-ad8da8ea1549 | /documentation/tutorials-operations/data-streaming-kafka-qdrant/ | Kafka Streaming into Qdrant | Installation of Qdrant-Kafka Sink Connector: | documentation | true | Kafka Streaming into Qdrant > Installation of Qdrant-Kafka Sink Connector:
To install the Qdrant Kafka connector using [Confluent Hub](https://www.confluent.io/hub/), you can utilize the straightforward `confluent-hub install` command. This command simplifies the process by eliminating the need for manual configuratio... | [
0.04257681220769882,
-0.07633757591247559,
-0.05680977925658226,
-0.037083614617586136,
0.019405238330364227,
-0.020805247128009796,
0.07343872636556625,
0.06867127865552902,
0.0745864287018776,
-0.01912834122776985,
0.023766111582517624,
0.011404975317418575,
0.04191373661160469,
-0.03366... |
f82e1213-2798-5842-9b01-b72179f1ab67 | /documentation/security/ | Security | Authentication > Table of Access | documentation | false | Security > Authentication > Table of Access
| Action | manage | read-only | collection read-write | collection read-only |
|--------|--------|-----------|----------------------|-----------------------|
| recommend groups | ✅ | ✅ | ✅ | ✅ |
| discover points | ✅ | ✅ | ✅ | ✅ |
| count points | ✅ | ✅ | ✅ | ✅ |
| version |... | [
0.005671835504472256,
-0.022200634703040123,
-0.030211610719561577,
-0.015386823564767838,
0.06166859716176987,
-0.010952367447316647,
0.022296635434031487,
0.08005349338054657,
0.05302619934082031,
-0.03540567681193352,
0.06287451833486557,
0.03730086237192154,
0.028440605849027634,
-0.03... |
f830b76a-0f59-5672-a492-147b4c35e2f7 | /course/essentials/day-3/sparse-retrieval-demo/ | Demo: Keyword Search with Sparse Vectors | Best Matching 25 (BM25) > BM25 in Qdrant > Create & Insert BM25-based Sparse Vectors | course | true | Demo: Keyword Search with Sparse Vectors > Best Matching 25 (BM25) > BM25 in Qdrant > Create & Insert BM25-based Sparse Vectors
This leaves us with the following `values` of the documents' words:
$$
\text{BM25}(d_i) = \mathrm{function}\left(\mathrm{TF}(d_i, D), k_1, b, |D|, \mathrm{avg}_{\text{corpus}}|D|\right)
$$
... | [
0.06986434757709503,
-0.038387481123209,
-0.028894823044538498,
-0.04236067831516266,
0.07631125301122665,
-0.04405072703957558,
0.06448360532522202,
0.03952083736658096,
0.05734007805585861,
-0.023146068677306175,
0.040449269115924835,
0.01870976947247982,
0.03925863280892372,
-0.03662052... |
f832a018-a5ce-5c6b-873d-7bf6a57c7dcb | /documentation/production-operations/multitenant-search/ | How to Implement Multitenancy and Custom Sharding in Qdrant | Before I go any further.....how secure is my user data? | documentation | false | How to Implement Multitenancy and Custom Sharding in Qdrant > Before I go any further.....how secure is my user data?
By design, Qdrant offers three levels of isolation. We initially introduced collection-based isolation, but your scaled setup has to move beyond this level. In this scenario, you will leverage payload-... | [
0.020306577906012535,
-0.007562791462987661,
-0.07113993167877197,
-0.003934474196285009,
0.05743921175599098,
-0.050151608884334564,
0.08796992152929306,
0.06771689653396606,
0.061651378870010376,
-0.02821207419037819,
0.007749759592115879,
0.036902107298374176,
0.06744939088821411,
-0.01... |
f84fd6f5-1446-507c-91ec-a5bb75f10523 | /documentation/overview/what-is-qdrant/ | What is Qdrant? | documentation | false | What is Qdrant?
Vector databases are a relatively new way for interacting with abstract data representations
derived from opaque machine learning models such as deep learning architectures. These
representations are often called vectors or embeddings and they are a compressed version of
the data used to train a mac... | [
0.049333952367305756,
-0.032510239630937576,
-0.056027524173259735,
-0.06468015164136887,
0.03696361184120178,
-0.032058678567409515,
0.10981841385364532,
0.06421082466840744,
0.050299063324928284,
-0.03133848309516907,
0.03418850526213646,
0.030505210161209106,
0.04613959789276123,
-0.050... | |
f87b454c-f698-56a2-b0f5-2f365ac3e317 | /course/essentials/day-6/congratulations/ | Course Completion and Next Steps | Your Learning Journey | course | false | Course Completion and Next Steps > Your Learning Journey
You've progressed from vector search fundamentals to production-ready expertise:
**Foundation Building** (Days 0-2): You mastered the core concepts of vector search, learned how similarity metrics work, and understood how HNSW indexing enables fast retrieval at... | [
0.024350324645638466,
-0.04151402413845062,
-0.026051869615912437,
0.0011537497630342841,
0.07108473032712936,
-0.056386422365903854,
0.034317370504140854,
0.049352191388607025,
0.07708075642585754,
-0.030651763081550598,
0.026236385107040405,
0.023807724937796593,
0.051994677633047104,
-0... |
f890cfa2-0cf3-5191-ac28-26754ab55561 | /course/essentials/day-3/hybrid-search/ | Hybrid Search and the Universal Query API | Key Takeaways | course | false | Hybrid Search and the Universal Query API > Key Takeaways
**What you've learned:**
- Dense vectors excel at semantic understanding and multilingual search
- Sparse vectors are ideal for exact matching and keyword-based retrieval
- Sequential retrieval + reranking has limitations (the first stage determines what's avai... | [
0.016145890578627586,
-0.025531215593218803,
-0.07836804538965225,
-0.04244300723075867,
0.06520938873291016,
-0.0425005704164505,
0.05193226784467697,
0.03281262516975403,
0.07750289142131805,
-0.04366866871714592,
0.036514248698949814,
0.03123423270881176,
0.061439938843250275,
-0.017172... |
f8954228-6d2e-559a-bc58-f8fb48740d21 | /documentation/tutorials-search-engineering/sparse-embeddings-ecommerce-part-1/ | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 1: Why Sparse Embeddings Beat BM25 | Why Qdrant for Sparse Vectors? | documentation | true | Fine-Tuning Sparse Embeddings for E-Commerce Search | Part 1: Why Sparse Embeddings Beat BM25 > Why Qdrant for Sparse Vectors?
Not every vector database treats sparse vectors as a first-class citizen. Qdrant does, and the difference matters in practice.
**Weighted sparse vectors.** SPLADE emits arbitrary learned weig... | [
0.04550128057599068,
-0.04596135392785072,
-0.02856115810573101,
-0.04768574982881546,
0.04622574523091316,
-0.049408961087465286,
0.08215520530939102,
0.020765259861946106,
0.06357195228338242,
-0.02325756475329399,
0.030487624928355217,
0.026343367993831635,
0.025006432086229324,
-0.0158... |
f899d760-a6b5-5f38-8305-068bd60d7f92 | /course/essentials/day-3/hybrid-search-demo/ | Demo: Implementing a Hybrid Search System | What You'll Discover | course | false | Demo: Implementing a Hybrid Search System > What You'll Discover
In the previous lesson, you learned the theory behind hybrid search and the Universal Query API. Today you'll implement it hands-on with a real dataset, comparing dense and sparse vector search and combining them using fusion algorithms.
**You'll learn ... | [
0.047891996800899506,
-0.02575937658548355,
-0.06550514698028564,
-0.033992040902376175,
0.06964829564094543,
-0.04374037683010101,
0.041170861572027206,
0.04309340566396713,
0.06779633462429047,
-0.024080222472548485,
0.04443732649087906,
0.05050910264253616,
0.06134217232465744,
-0.01134... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.