pgvector vs Qdrant MCP Server: both are coding tools on Scriptiii AI Tools. pgvector pricing: Free. Free to use. (reported by the maker) Qdrant MCP Server pricing: Free. Free to use. (reported by the maker) Both have 0 upvotes.
| Field | pgvector | Qdrant MCP Server |
|---|---|---|
| Page | pgvector | Qdrant MCP Server |
| Pricing (maker-reported) | Free. Free to use. | Free. Free to use. |
| Upvotes | 0 | 0 |
| Best for | ai-agents, tool, vector-databases | ai-agents, mcp-server, vector-databases |
| Tagline | Postgres extension for vector similarity search. | Official Qdrant MCP server that stores and retrieves memories in a Qdrant vector database. |
| Website | Visit pgvector | Visit Qdrant MCP Server |
Pricing and pricing details come from the maker's submission. They are not independently verified.
Open-source vector similarity search for Postgres
Store your vectors with the rest of your data. Supports:
Plus ACID compliance, point-in-time recovery, JOINs, and all of the other great features of Postgres
Have a lot of vectors? Use quantization to scale
Compile and install the extension (supports Postgres 13+)
cd /tmp
git clone --branch v0.8.7 https://github.com/pgvector/pgvector.git
cd pgvector
make
make install # may need sudo
See the installation notes if you run into issues
You can also install it with Docker, Homebrew, PGXN, APT, Yum, pkg, APK, or conda-forge, and it comes preinstalled with Postgres.app and many hosted providers. There are also instructions for GitHub Actions.
Ensure C++ support in Visual Studio is installed and run x64 Native Tools Command Prompt for VS [version] as administrator. Then use nmake to build:
set "PGROOT=C:\Program Files\PostgreSQL\18"
cd %TEMP%
git clone --branch v0.8.7 https://github.com/pgvector/pgvector.git
cd pgvector
nmake /F Makefile.win
nmake /F Makefile.win install
See the installation notes if you run into issues
You can also install it with Docker or conda-forge.
Enable the extension (do this once in each database where you want to use it)
CREATE EXTENSION vector;
Create a vector column with 3 dimensions
CREATE TABLE items (id bigserial PRIMARY KEY, embedding vector(3));
Insert vectors
INSERT INTO items (embedding) VALUES ('[1,2,3]'), ('[4,5,6]');
Get the nearest neighbors by L2 distance
SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5;
Also supports inner product (<#>), cosine distance (<=>), and L1 distance (<+>)
Note: <#> returns the negative inner product since Postgres only supports ASC order index scans on operators
Create a new table with a vector column
CREATE TABLE items (id bigserial PRIMARY KEY, embedding vector(3));
Or add a vector column to an existing table
ALTER TABLE items ADD COLUMN embedding vector(3);
Also supports half-precision, binary, and sparse vectors
Insert vectors
INSERT INTO items (embedding) VALUES ('[1,2,3]'), ('[4,5,6]');
Or load vectors in bulk using COPY (example)
COPY items (embedding) FROM STDIN WITH (FORMAT BINARY);
Upsert vectors
INSERT INTO items (id, embedding) VALUES (1, '[1,2,3]'), (2, '[4,5,6]')
ON CONFLICT (id) DO UPDATE SET embedding = EXCLUDED.embedding;
Update vectors
UPDATE items SET embedding = '[1,2,3]' WHERE id = 1;
Delete vectors
DELETE FROM items WHERE id = 1;
Get the nearest neighbors to a vector
SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5;
Supported distance functions are:
<-> - L2 distance<#> - (negative) inner product<=> - cosine distance<+> - L1 distance<~> - Hamming distance (binary vectors)<%> - Jaccard distance (binary vectors)Get the nearest neighbors to a row
SELECT * FROM items WHERE id != 1 ORDER BY embedding <-> (SELECT embedding FROM items WHERE id = 1) LIMIT 5;
Get rows within a certain distance
SELECT * FROM items WHERE embedding <-> '[3,1,2]' < 5;
Note: Combine with ORDER BY and LIMIT to use an index
Get the distance
SELECT embedding <-> '[3,1,2]' AS distance FROM items;
For inner product, multiply by -1 (since <#> returns the negative inner product)
SELECT (embedding <#> '[3,1,2]') * -1 AS inner_product FROM items;
For cosine similarity, use 1 - cosine distance
SELECT 1 - (embedding <=> '[3,1,2]') AS cosine_similarity FROM items;
Average vectors
SELECT AVG(embedding) FROM items;
Average groups of vectors
SELECT category_id, AVG(embedding) FROM items GROUP BY category_id;
By default, pgvector performs exact nearest neighbor search, which provides perfect recall.
You can add an index to use approximate nearest neighbor search, which trades some recall for speed. Unlike typical indexes, you will see different results for queries after adding an approximate index.
Supported index types are:
An HNSW index creates a multilayer graph. It has better query performance than IVFFlat (in terms of speed-recall tradeoff), but has slower build times and uses more memory. Also, an index can be created without any data in the table since there isn’t a training step like IVFFlat.
Add an index for each distance function you want to use.
L2 distance
CREATE INDEX ON items USING hnsw (embedding vector_l2_ops);
Note: Use halfvec_l2_ops for halfvec and sparsevec_l2_ops for sparsevec (and similar with the other distance functions)
Inner product
CREATE INDEX ON items USING hnsw (embedding vector_ip_ops);
Cosine distance
CREATE INDEX ON items USING hnsw (embedding vector_cosine_ops);
L1 distance
CREATE INDEX ON items USING hnsw (embedding vector_l1_ops);
Hamming distance
CREATE INDEX ON items USING hnsw (embedding bit_hamming_ops);
Jaccard distance
CREATE INDEX ON items USING hnsw (embedding bit_jaccard_ops);
Supported types are:
vector - up to 2,000 dimensionshalfvec - up to 4,000 dimensionsbit - up to 64,000 dimensionssparsevec - up to 1,000 non-zero elementsSpecify HNSW parameters
m - the max number of connections per layer (16 by default)ef_construction - the size of the dynamic candidate list for constructing the graph (64 by default)CREATE INDEX ON items USING hnsw (embedding vector_l2_ops) WITH (m = 16, ef_construction = 64);
A higher value of ef_construction provides better recall at the cost of index build time / insert speed.
Specify the size of the dynamic candidate list for search (40 by default)
SET hnsw.ef_search = 100;
A higher value provides better recall at the cost of speed.
Use SET LOCAL inside a transaction to set it for a single query
BEGIN;
SET LOCAL hnsw.ef_search = 100;
SELECT ...
COMMIT;
Indexes build significantly faster when the graph fits into maintenance_work_mem
SET maintenance_work_mem = '8GB';
A notice is shown when the graph no longer fits
NOTICE: hnsw graph no longer fits into maintenance_work_mem after 100000 tuples
DETAIL: Building will take significantly more time.
HINT: Increase maintenance_work_mem to speed up builds.
Note: Do not set maintenance_work_mem so high that it exhausts the memory on the server
Like other index types, it’s faster to create an index after loading your initial data
You can also speed up index creation by increasing the number of parallel workers (2 by default)
SET max_parallel_maintenance_workers = 7; -- plus leader
For a large number of workers, you may need to increase max_parallel_workers (8 by default)
The index options also have a significant impact on build time (use the defaults unless seeing low recall)
Use binary quantization for faster build times at scale
Check indexing progress
SELECT phase, round(100.0 * blocks_done / nullif(blocks_total, 0), 1) AS "%" FROM pg_stat_progress_create_index;
The phases for HNSW are:
initializingloading tuplesAn IVFFlat index divides vectors into lists, and then searches a subset of those lists that are closest to the query vector. It has faster build times and uses less memory than HNSW, but has lower query performance (in terms of speed-recall tradeoff).
Three keys to achieving good recall are:
rows / 1000 for up to 1M rows and sqrt(rows) for over 1M rowssqrt(lists)Add an index for each distance function you want to use.
L2 distance
CREATE INDEX ON items USING ivfflat (embedding vector_l2_ops) WITH (lists = 100);
Note: Use halfvec_l2_ops for halfvec (and similar with the other distance functions)
Inner product
CREATE INDEX ON items USING ivfflat (embedding vector_ip_ops) WITH (lists = 100);
Cosine distance
CREATE INDEX ON items USING ivfflat (embedding vector_cosine_ops) WITH (lists = 100);
Hamming distance
CREATE INDEX ON items USING ivfflat (embedding bit_hamming_ops) WITH (lists = 100);
Supported types are:
vector - up to 2,000 dimensionshalfvec - up to 4,000 dimensionsbit - up to 64,000 dimensionsSpecify the number of probes (1 by default)
SET ivfflat.probes = 10;
A higher value provides better recall at the cost of speed, and it can be set to the number of lists for exact nearest neighbor search (at which point the planner won’t use the index)
Use SET LOCAL inside a transaction to set it for a single query
BEGIN;
SET LOCAL ivfflat.probes = 10;
SELECT ...
COMMIT;
Speed up index creation on large tables by increasing the number of parallel workers (2 by default)
SET max_parallel_maintenance_workers = 7; -- plus leader
For a large number of workers, you may also need to increase max_parallel_workers (8 by default)
Check indexing progress
SELECT phase, round(100.0 * tuples_done / nullif(tuples_total, 0), 1) AS "%" FROM pg_stat_progress_create_index;
The phases for IVFFlat are:
initializingperforming k-meansassigning tuplesloading tuplesNote: % is only populated during the loading tuples phase
There are a few ways to index nearest neighbor queries with a WHERE clause.
SELECT * FROM items WHERE category_id = 123 ORDER BY embedding <-> '[3,1,2]' LIMIT 5;
A good place to start is creating an index on the filter column. This can provide fast, exact nearest neighbor search in many cases. Postgres has a number of index types for this: B-tree (default), hash, GiST, SP-GiST, GIN, and BRIN.
CREATE INDEX ON items (category_id);
For multiple columns, consider a multicolumn index.
CREATE INDEX ON items (location_id, category_id);
Exact indexes work well for conditions that match a low percentage of rows. Otherwise, approximate indexes can work better.
CREATE INDEX ON items USING hnsw (embedding vector_l2_ops);
With approximate indexes, filtering is applied after the index is scanned. If a condition matches 10% of rows, with HNSW and the default hnsw.ef_search of 40, only 4 rows will match on average. For more rows, enable iterative index scans, which will automatically scan more of the index when needed.
SET hnsw.iterative_scan = strict_order;
If filtering by only a few distinct values, consider partial indexing.
CREATE INDEX ON items USING hnsw (embedding vector_l2_ops) WHERE (category_id = 123);
If filtering by many different values, consider partitioning.
CREATE TABLE items (embedding vector(3), category_id int) PARTITION BY LIST(category_id);
For applications with multiple tenants, sharing an approximate index between tenants means vectors from one tenant can affect recall (and speed) for other tenants.
For tenant isolation, use list partitioning or separate tables.
CREATE TABLE items (customer_id int, embedding vector(3)) PARTITION BY LIST(customer_id);
With approximate indexes, queries with filtering can return less results since filtering is applied after the index is scanned. Starting with 0.8.0, you can enable iterative index scans, which will automatically scan more of the index until enough results are found (or it reaches hnsw.max_scan_tuples or ivfflat.max_probes).
Iterative scans can use strict or relaxed ordering.
Strict ensures results are in the exact order by distance
SET hnsw.iterative_scan = strict_order;
Relaxed allows results to be slightly out of order by distance, but provides better recall
SET hnsw.iterative_scan = relaxed_order;
## or
SET ivfflat.iterative_scan = relaxed_order;
With relaxed ordering, you can use a materialized CTE to get strict ordering
WITH relaxed_results AS MATERIALIZED (
SELECT id, embedding <-> '[1,2,3]' AS distance FROM items WHERE category_id = 123 ORDER BY distance LIMIT 5
) SELECT * FROM relaxed_results ORDER BY distance + 0;
Note: + 0 is needed for Postgres 17+
For queries that filter by distance, use a materialized CTE and place the distance filter outside of it for best performance (due to the current behavior of the Postgres executor)
WITH nearest_results AS MATERIALIZED (
SELECT id, embedding <-> '[1,2,3]' AS distance FROM items ORDER BY distance LIMIT 5
) SELECT * FROM nearest_results WHERE distance < 5 ORDER BY distance;
Note: Place any other filters inside the CTE
Since scanning a large portion of an approximate index is expensive, there are options to control when a scan ends.
Specify the max number of tuples to visit (20,000 by default)
SET hnsw.max_scan_tuples = 20000;
Note: This is approximate and does not affect the initial scan
Specify the max amount of memory to use, as a multiple of work_mem (1 by default)
SET hnsw.scan_mem_multiplier = 2;
Note: Try increasing this if increasing hnsw.max_scan_tuples does not improve recall
Specify the max number of probes
SET ivfflat.max_probes = 100;
Note: If this is lower than ivfflat.probes, ivfflat.probes will be used
Use the halfvec type to store half-precision vectors
CREATE TABLE items (id bigserial PRIMARY KEY, embedding halfvec(3));
Index vectors at half precision for smaller indexes
CREATE INDEX ON items USING hnsw ((embedding::halfvec(3)) halfvec_l2_ops);
Get the nearest neighbors
SELECT * FROM items ORDER BY embedding::halfvec(3) <-> '[1,2,3]' LIMIT 5;
Use the bit type to store binary vectors (example)
CREATE TABLE items (id bigserial PRIMARY KEY, embedding bit(3));
INSERT INTO items (embedding) VALUES ('000'), ('111');
Get the nearest neighbors by Hamming distance
SELECT * FROM items ORDER BY embedding <~> '101' LIMIT 5;
Also supports Jaccard distance (<%>)
Use expression indexing for binary quantization
CREATE INDEX ON items USING hnsw ((binary_quantize(embedding)::bit(3)) bit_hamming_ops);
Get the nearest neighbors by Hamming distance
SELECT * FROM items ORDER BY binary_quantize(embedding)::bit(3) <~> binary_quantize('[1,-2,3]') LIMIT 5;
Re-rank by the original vectors for better recall
SELECT * FROM (
SELECT * FROM items ORDER BY binary_quantize(embedding)::bit(3) <~> binary_quantize('[1,-2,3]') LIMIT 20
) ORDER BY embedding <=> '[1,-2,3]' LIMIT 5;
Use the sparsevec type to store sparse vectors
CREATE TABLE items (id bigserial PRIMARY KEY, embedding sparsevec(5));
Insert vectors
INSERT INTO items (embedding) VALUES ('{1:1,3:2,5:3}/5'), ('{1:4,3:5,5:6}/5');
The format is {index1:value1,index2:value2}/dimensions and indices start at 1 like SQL arrays
Get the nearest neighbors by L2 distance
SELECT * FROM items ORDER BY embedding <-> '{1:3,3:1,5:2}/5' LIMIT 5;
Use together with Postgres full-text search for hybrid search.
SELECT id, content FROM items, plainto_tsquery('hello search') query
WHERE textsearch @@ query ORDER BY ts_rank_cd(textsearch, query) DESC LIMIT 5;
You can use Reciprocal Rank Fusion or a cross-encoder to combine results.
Use expression indexing to index subvectors
CREATE INDEX ON items USING hnsw ((subvector(embedding, 1, 3)::vector(3)) vector_cosine_ops);
Get the nearest neighbors by cosine distance
SELECT * FROM items ORDER BY subvector(embedding, 1, 3)::vector(3) <=> subvector('[1,2,3,4,5]'::vector, 1, 3) LIMIT 5;
Re-rank by the full vectors for better recall
SELECT * FROM (
SELECT * FROM items ORDER BY subvector(embedding, 1, 3)::vector(3) <=> subvector('[1,2,3,4,5]'::vector, 1, 3) LIMIT 20
) ORDER BY embedding <=> '[1,2,3,4,5]' LIMIT 5;
Use a tool like PgTune to set initial values for Postgres server parameters. For instance, shared_buffers should typically be 25% of the server’s memory. You can find the config file with:
SHOW config_file;
And check individual settings with:
SHOW shared_buffers;
Be sure to restart Postgres for changes to take effect.
Use the halfvec type instead of vector for a smaller working set.
Use COPY for bulk loading data (example).
COPY items (embedding) FROM STDIN WITH (FORMAT BINARY);
Add any indexes after loading the initial data for best performance.
See index build time for HNSW and IVFFlat.
Use binary quantization for smaller indexes and faster build times at scale.
In production environments, create indexes concurrently to avoid blocking writes.
CREATE INDEX CONCURRENTLY ...
Use EXPLAIN (ANALYZE, BUFFERS) to debug performance.
EXPLAIN (ANALYZE, BUFFERS) SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5;
To speed up queries without an index, increase max_parallel_workers_per_gather.
SET max_parallel_workers_per_gather = 4;
If vectors are normalized to length 1 (like OpenAI embeddings), use inner product for best performance.
SELECT * FROM items ORDER BY embedding <#> '[3,1,2]' LIMIT 5;
Use binary quantization with re-ranking to keep indexes in-memory at scale.
To speed up queries with an IVFFlat index, increase the number of inverted lists (at the expense of recall).
CREATE INDEX ON items USING ivfflat (embedding vector_l2_ops) WITH (lists = 1000);
Vacuuming can take a while for HNSW indexes. Speed it up by reindexing first.
REINDEX INDEX CONCURRENTLY index_name;
VACUUM table_name;
For a smaller working set:
halfvec type instead of vector for tablesScale vertically by increasing memory, CPU, and storage on a single instance. Use existing tools to tune parameters and monitor performance.
Scale horizontally with replicas, or use Citus, PgDog, or another approach for sharding (example).
Use existing tools like pg_stat_statements or PgHero to monitor performance.
Monitor recall by comparing results from approximate search with exact search.
BEGIN;
SET LOCAL enable_indexscan = off; -- use exact search
SELECT ...
COMMIT;
Use pgvector from any language with a Postgres client. You can even generate and store vectors in one language and query them in another.
| Language | Libraries / Examples |
|---|---|
| Ada | pgvector-ada |
| Algol | pgvector-algol |
| C | pgvector-c |
| C++ | pgvector-cpp |
| C#, F#, Visual Basic | pgvector-dotnet |
| COBOL | pgvector-cobol |
| Crystal | pgvector-crystal |
| D | pgvector-d |
| Dart | pgvector-dart |
| Elixir | pgvector-elixir |
| Erlang | pgvector-erlang |
| Fortran | pgvector-fortran |
| Gleam | pgvector-gleam |
| Go | pgvector-go |
| Haskell | pgvector-haskell |
| Java, Kotlin, Groovy, Scala |
A non-partitioned table has a limit of 32 TB by default in Postgres. A partitioned table can have thousands of partitions of that size.
Yes, pgvector uses the write-ahead log (WAL), which allows for replication and point-in-time recovery.
You can use half-precision vectors or half-precision indexing to index up to 4,000 dimensions or binary quantization to index up to 64,000 dimensions. Other options are indexing subvectors (for models that support it) or dimensionality reduction.
You can use vector as the type (instead of vector(n)).
CREATE TABLE embeddings (model_id bigint, item_id bigint, embedding vector, PRIMARY KEY (model_id, item_id));
However, you can only create indexes on rows with the same number of dimensions (using expression and partial indexing):
CREATE INDEX ON embeddings USING hnsw ((embedding::vector(3)) vector_l2_ops) WHERE (model_id = 123);
and query with:
SELECT * FROM embeddings WHERE model_id = 123 ORDER BY embedding::vector(3) <-> '[3,1,2]' LIMIT 5;
You can use the double precision[] or numeric[] type to store vectors with more precision.
CREATE TABLE items (id bigserial PRIMARY KEY, embedding double precision[]);
-- use {} instead of [] for Postgres arrays
INSERT INTO items (embedding) VALUES ('{1,2,3}'), ('{4,5,6}');
Optionally, add a check constraint to ensure data can be converted to the vector type and has the expected dimensions.
ALTER TABLE items ADD CHECK (vector_dims(embedding::vector) = 3);
Use expression indexing to index (at a lower precision):
CREATE INDEX ON items USING hnsw ((embedding::vector(3)) vector_l2_ops);
and query with:
SELECT * FROM items ORDER BY embedding::vector(3) <-> '[3,1,2]' LIMIT 5;
No, but like other index types, you’ll likely see better performance if they do. You can get the size of an index with:
SELECT pg_size_pretty(pg_relation_size('index_name'));
Use half-precision indexing or binary quantization for smaller indexes.
The query needs to have an ORDER BY and LIMIT, and the ORDER BY must be the result of a distance operator (not an expression) in ascending order.
-- index
ORDER BY embedding <=> '[3,1,2]' LIMIT 5;
-- no index
ORDER BY 1 - (embedding <=> '[3,1,2]') DESC LIMIT 5;
You can encourage the planner to use an index for a query with:
BEGIN;
SET LOCAL enable_seqscan = off;
SELECT ...
COMMIT;
Also, if the table is small, a table scan may be faster.
The planner doesn’t consider out-of-line storage in cost estimates, which can make a serial scan look cheaper. You can reduce the cost of a parallel scan for a query with:
BEGIN;
SET LOCAL min_parallel_table_scan_size = 1;
SET LOCAL parallel_setup_cost = 1;
SELECT ...
COMMIT;
or choose to store vectors inline:
ALTER TABLE items ALTER COLUMN embedding SET STORAGE PLAIN;
Results are limited by the size of the dynamic candidate list (hnsw.ef_search), which is 40 by default. There may be even less results due to dead tuples or filtering conditions in the query. Enabling iterative index scans can help address this.
Also, note that NULL vectors are not indexed (as well as zero vectors for cosine distance).
The index was likely created with too little data for the number of lists. Drop the index until the table has more data.
DROP INDEX index_name;
Results can also be limited by the number of probes (ivfflat.probes). Enabling iterative index scans can address this.
Also, note that NULL vectors are not indexed (as well as zero vectors for cosine distance).
Each vector takes 4 * dimensions + 8 bytes of storage. Each element is a single-precision floating-point number (like the real type in Postgres), and all elements must be finite (no NaN, Infinity or -Infinity). Vectors can have up to 16,000 dimensions.
| Operator | Description | Added |
|---|---|---|
| + | element-wise addition | |
| - | element-wise subtraction | |
| * | element-wise multiplication | 0.5.0 |
| || | concatenate | 0.7.0 |
| <-> | Euclidean distance | |
| <#> | negative inner product | |
| <=> | cosine distance | |
| <+> | taxicab distance | 0.7.0 |
| Function | Description | Added |
|---|---|---|
| binary_quantize(vector) → bit | binary quantize | 0.7.0 |
| cosine_distance(vector, vector) → double precision | cosine distance | |
| inner_product(vector, vector) → double precision | inner product | |
| l1_distance(vector, vector) → double precision | taxicab distance | 0.5.0 |
| l2_distance(vector, vector) → double precision | Euclidean distance | |
| l2_normalize(vector) → vector | normalize with Euclidean norm | 0.7.0 |
| subvector(vector, integer, integer) → vector | subvector | 0.7.0 |
| vector_dims(vector) → integer | number of dimensions | |
| vector_norm(vector) → double precision | Euclidean norm |
| Function | Description | Added |
|---|---|---|
| avg(vector) → vector | average | |
| sum(vector) → vector | sum | 0.5.0 |
Each half vector takes 2 * dimensions + 8 bytes of storage. Each element is a half-precision floating-point number, and all elements must be finite (no NaN, Infinity or -Infinity). Half vectors can have up to 16,000 dimensions.
| Operator | Description | Added |
|---|---|---|
| + | element-wise addition | 0.7.0 |
| - | element-wise subtraction | 0.7.0 |
| * | element-wise multiplication | 0.7.0 |
| || | concatenate | 0.7.0 |
| <-> | Euclidean distance | 0.7.0 |
| <#> | negative inner product | 0.7.0 |
| <=> | cosine distance | 0.7.0 |
| <+> | taxicab distance | 0.7.0 |
| Function | Description | Added |
|---|---|---|
| binary_quantize(halfvec) → bit | binary quantize | 0.7.0 |
| cosine_distance(halfvec, halfvec) → double precision | cosine distance | 0.7.0 |
| inner_product(halfvec, halfvec) → double precision | inner product | 0.7.0 |
| l1_distance(halfvec, halfvec) → double precision | taxicab distance | 0.7.0 |
| l2_distance(halfvec, halfvec) → double precision | Euclidean distance | 0.7.0 |
| l2_norm(halfvec) → double precision | Euclidean norm | 0.7.0 |
| l2_normalize(halfvec) → halfvec | normalize with Euclidean norm | 0.7.0 |
| subvector(halfvec, integer, integer) → halfvec | subvector | 0.7.0 |
| vector_dims(halfvec) → integer | number of dimensions | 0.7.0 |
| Function | Description | Added |
|---|---|---|
| avg(halfvec) → halfvec | average | 0.7.0 |
| sum(halfvec) → halfvec | sum | 0.7.0 |
Each bit vector takes dimensions / 8 + 8 bytes of storage. See the Postgres docs for more info.
| Operator | Description | Added |
|---|---|---|
| <~> | Hamming distance | 0.7.0 |
| <%> | Jaccard distance | 0.7.0 |
| Function | Description | Added |
|---|---|---|
| hamming_distance(bit, bit) → double precision | Hamming distance | 0.7.0 |
| jaccard_distance(bit, bit) → double precision | Jaccard distance | 0.7.0 |
Each sparse vector takes 8 * non-zero elements + 16 bytes of storage. Each element is a single-precision floating-point number, and all elements must be finite (no NaN, Infinity or -Infinity). Sparse vectors can have up to 16,000 non-zero elements.
| Operator | Description | Added |
|---|---|---|
| <-> | Euclidean distance | 0.7.0 |
| <#> | negative inner product | 0.7.0 |
| <=> | cosine distance | 0.7.0 |
| <+> | taxicab distance | 0.7.0 |
| Function | Description | Added |
|---|---|---|
| cosine_distance(sparsevec, sparsevec) → double precision | cosine distance | 0.7.0 |
| inner_product(sparsevec, sparsevec) → double precision | inner product | 0.7.0 |
| l1_distance(sparsevec, sparsevec) → double precision | taxicab distance | 0.7.0 |
| l2_distance(sparsevec, sparsevec) → double precision | Euclidean distance | 0.7.0 |
| l2_norm(sparsevec) → double precision | Euclidean norm | 0.7.0 |
| l2_normalize(sparsevec) → sparsevec | normalize with Euclidean norm | 0.7.0 |
If your machine has multiple Postgres installations, specify the path to pg_config with:
export PG_CONFIG=/Library/PostgreSQL/18/bin/pg_config
Then re-run the installation instructions (run make clean before make if needed). If sudo is needed for make install, use:
sudo --preserve-env=PG_CONFIG make install
A few common paths on Mac are:
/Library/PostgreSQL/18/bin/pg_config/opt/homebrew/opt/postgresql@18/bin/pg_config/usr/local/opt/postgresql@18/bin/pg_configNote: Replace 18 with your Postgres server version
If compilation fails with fatal error: postgres.h: No such file or directory, make sure Postgres development files are installed on the server.
For Ubuntu and Debian, use:
sudo apt install postgresql-server-dev-18
Note: Replace 18 with your Postgres server version
If compilation fails and the output includes warning: no such sysroot directory on Mac, your Postgres installation points to a path that no longer exists.
pg_config --cppflags
Reinstall Postgres to fix this.
By default, pgvector compiles with -march=native on some platforms for best performance. However, this can lead to Illegal instruction errors if trying to run the compiled extension on a different machine.
To compile for portability, use:
make OPTFLAGS=""
If compilation fails with Cannot open include file: 'postgres.h': No such file or directory, make sure PGROOT is correct.
If compilation fails with error C2196: case value '4' already used, make sure you’re using the x64 Native Tools Command Prompt. Then run nmake /F Makefile.win clean and re-run the installation instructions.
If linking fails with unresolved external symbol float_to_shortest_decimal_bufn with Postgres 17.0-17.2, upgrade to Postgres 17.3+.
If installation fails with Access is denied, re-run the installation instructions as an administrator.
Get the Docker image with:
docker pull pgvector/pgvector:pg18-trixie
This adds pgvector to the Postgres image (replace 18 with your Postgres server version, and run it the same way).
Supported tags are:
pg18-trixie, 0.8.7-pg18-trixiepg18-bookworm, 0.8.7-pg18-bookworm, pg18, 0.8.7-pg18pg17-trixie, 0.8.7-pg17-trixiepg17-bookworm, 0.8.7-pg17-bookworm, pg17, 0.8.7-pg17pg16-trixie, 0.8.7-pg16-trixiepg16-bookworm, 0.8.7-pg16-bookworm, pg16, 0.8.7-pg16pg15-trixie, 0.8.7-pg15-trixiepg15-bookworm, 0.8.7-pg15-bookworm, pg15, 0.8.7-pg15pg14-trixie, 0.8.7-pg14-trixiepg14-bookworm, 0.8.7-pg14-bookworm, pg14, 0.8.7-pg14pg13-trixie, 0.8.7-pg13-trixiepg13-bookworm, 0.8.7-pg13-bookworm, pg13, 0.8.7-pg13You can also build the image manually:
git clone --branch v0.8.7 https://github.com/pgvector/pgvector.git
cd pgvector
docker build --pull --build-arg PG_MAJOR=18 -t myuser/pgvector .
If you increase maintenance_work_mem, make sure --shm-size is at least that size to avoid an error with parallel HNSW index builds.
docker run --shm-size=1g ...
With Homebrew Postgres, you can use:
brew install pgvector
Note: This only adds it to the postgresql@18 and postgresql@17 formulas
Install from the PostgreSQL Extension Network with:
pgxn install vector
Debian and Ubuntu packages are available from the PostgreSQL APT Repository. Follow the setup instructions and run:
sudo apt install postgresql-18-pgvector
Note: Replace 18 with your Postgres server version
RPM packages are available from the PostgreSQL Yum Repository. Follow the setup instructions for your distribution and run:
sudo yum install pgvector_18
## or
sudo dnf install pgvector_18
Note: Replace 18 with your Postgres server version
Install the FreeBSD package with:
pkg install postgresql18-pgvector
or the port with:
cd /usr/ports/databases/pgvector
make install
Install the Alpine package with:
apk add postgresql-pgvector
With Conda Postgres, install from conda-forge with:
conda install -c conda-forge pgvector
This method is community-maintained by @mmcauliffe
Download the latest release with Postgres 15+.
pgvector is available on these providers.
Install the latest version (use the same method as the original installation). Then in each database you want to upgrade, run:
ALTER EXTENSION vector UPDATE;
You can check the version in the current database with:
SELECT extversion FROM pg_extension WHERE extname = 'vector';
Thanks to:
View the changelog
Everyone is encouraged to help improve this project. Here are a few ways you can help:
To get started with development:
git clone https://github.com/pgvector/pgvector.git
cd pgvector
make
make install
To run all tests:
make installcheck # regression tests
make prove_installcheck # TAP tests
To run single tests:
make installcheck REGRESS=functions # regression test
make prove_installcheck PROVE_TESTS=test/t/001_ivfflat_wal.pl # TAP test
To enable assertions:
make clean && PG_CFLAGS="-DUSE_ASSERT_CHECKING" make && make install
To enable benchmarking:
make clean && PG_CFLAGS="-DHNSW_BENCH -DIVFFLAT_BENCH" make && make install
To show memory usage:
make clean && PG_CFLAGS="-DHNSW_MEMORY -DIVFFLAT_MEMORY" make && make install
To get k-means metrics:
make clean && PG_CFLAGS="-DIVFFLAT_KMEANS_DEBUG" make && make install
Resources for contributors
The Model Context Protocol (MCP) is an open protocol that enables
seamless integration between LLM applications and external data sources and tools. Whether you're building an
AI-powered IDE, enhancing a chat interface, or creating custom AI workflows, MCP provides a standardized way to
connect LLMs with the context they need.
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
An official Model Context Protocol server for keeping and retrieving memories in the Qdrant vector search engine. It acts as a semantic memory layer on top of the Qdrant database.
| pgvector-java |
| JavaScript, TypeScript | pgvector-node |
| Julia | Pgvector.jl |
| Lisp | pgvector-lisp |
| Lua | pgvector-lua |
| Nim | pgvector-nim |
| OCaml | pgvector-ocaml |
| Pascal | pgvector-pascal |
| Perl | pgvector-perl |
| PHP | pgvector-php |
| Prolog | pgvector-prolog |
| Python | pgvector-python |
| R | pgvector-r |
| Racket | pgvector-racket |
| Raku | pgvector-raku |
| Ruby | pgvector-ruby, Neighbor |
| Rust | pgvector-rust |
| Swift | pgvector-swift |
| Tcl | pgvector-tcl |
| Zig | pgvector-zig |
qdrant-store
information (string): Information to storemetadata (JSON): Optional metadata to storecollection_name (string): Name of the collection to store the information in. This field is required if there are no default collection name.qdrant-find
query (string): Query to use for searchingcollection_name (string): Name of the collection to store the information in. This field is required if there are no default collection name.Configuration is done via environment variables. The only command-line argument is --transport, used to select the transport protocol.
[!NOTE]
You cannot provide bothQDRANT_URLandQDRANT_LOCAL_PATHat the same time.
| Name | Description | Default Value |
|---|---|---|
QDRANT_URL | URL of the Qdrant server | None |
QDRANT_API_KEY | API key for the Qdrant server | None |
COLLECTION_NAME | Name of the default collection to use. | None |
QDRANT_LOCAL_PATH | Path to the local Qdrant database (alternative to QDRANT_URL) | None |
EMBEDDING_PROVIDER | Embedding provider to use (currently only "fastembed" is supported) | fastembed |
EMBEDDING_MODEL | Name of the embedding model to use | sentence-transformers/all-MiniLM-L6-v2 |
TOOL_STORE_DESCRIPTION | Custom description for the store tool | See default in settings.py |
TOOL_FIND_DESCRIPTION | Custom description for the find tool | See default in settings.py |
QDRANT_SEARCH_LIMIT | Maximum number of results to return from search | 10 |
QDRANT_READ_ONLY | Enable read-only mode (disables qdrant-store tool) | false |
Since mcp-server-qdrant is based on FastMCP, it also supports all the FastMCP environment variables. The most
important ones are listed below:
| Environment Variable | Description | Default Value |
|---|---|---|
FASTMCP_LOG_LEVEL | Set logging level (DEBUG, INFO, WARNING, ERROR, CRITICAL) | INFO |
FASTMCP_DEBUG | Enable debug mode | false |
FASTMCP_HOST | Host address to bind the server to (SSE and HTTP transports) | 127.0.0.1 |
FASTMCP_PORT | Port to run the server on (SSE and HTTP transports) | 8000 |
FASTMCP_MASK_ERROR_DETAILS | Hide exception details from tool error messages sent to clients | false |
[!NOTE]
Before version 0.9.0, these settings used theFASTMCP_SERVER_prefix, for exampleFASTMCP_SERVER_HOST.
The oldFASTMCP_SERVER_HOST,FASTMCP_SERVER_PORT, andFASTMCP_SERVER_DEBUGnames still work but are deprecated
and will be removed in a future release. If both names are set, the new one wins.
When using uvx no specific installation is needed to directly run mcp-server-qdrant.
QDRANT_URL="http://localhost:6333" \
COLLECTION_NAME="my-collection" \
EMBEDDING_MODEL="sentence-transformers/all-MiniLM-L6-v2" \
uvx mcp-server-qdrant
The server supports different transport protocols that can be specified using the --transport flag:
QDRANT_URL="http://localhost:6333" \
COLLECTION_NAME="my-collection" \
uvx mcp-server-qdrant --transport sse
Supported transport protocols:
stdio (default): Standard input/output transport, might only be used by local MCP clientssse: Server-Sent Events transport, perfect for remote clientsstreamable-http: Streamable HTTP transport, perfect for remote clients, more recent than SSEThe default transport is stdio if not specified.
When SSE transport is used, the server will listen on the specified port and wait for incoming connections. The default
port is 8000, however it can be changed using the FASTMCP_PORT environment variable.
QDRANT_URL="http://localhost:6333" \
COLLECTION_NAME="my-collection" \
FASTMCP_PORT=1234 \
uvx mcp-server-qdrant --transport sse
A Dockerfile is available for building and running the MCP server:
## Build the container
docker build -t mcp-server-qdrant .
## Run the container
docker run -p 8000:8000 \
-e FASTMCP_HOST="0.0.0.0" \
-e QDRANT_URL="http://your-qdrant-server:6333" \
-e QDRANT_API_KEY="your-api-key" \
-e COLLECTION_NAME="your-collection" \
mcp-server-qdrant
[!TIP]
Please note that we setFASTMCP_HOST="0.0.0.0"to make the server listen on all network interfaces. This is
necessary when running the server in a Docker container.
To install Qdrant MCP Server for Claude Desktop automatically via Smithery:
npx @smithery/cli install mcp-server-qdrant --client claude
To use this server with the Claude Desktop app, add the following configuration to the "mcpServers" section of your
claude_desktop_config.json:
{
"qdrant": {
"command": "uvx",
"args": ["mcp-server-qdrant"],
"env": {
"QDRANT_URL": "https://xyz-example.eu-central.aws.cloud.qdrant.io:6333",
"QDRANT_API_KEY": "your_api_key",
"COLLECTION_NAME": "your-collection-name",
"EMBEDDING_MODEL": "sentence-transformers/all-MiniLM-L6-v2"
}
}
}
For local Qdrant mode:
{
"qdrant": {
"command": "uvx",
"args": ["mcp-server-qdrant"],
"env": {
"QDRANT_LOCAL_PATH": "/path/to/qdrant/database",
"COLLECTION_NAME": "your-collection-name",
"EMBEDDING_MODEL": "sentence-transformers/all-MiniLM-L6-v2"
}
}
}
This MCP server will automatically create a collection with the specified name if it doesn't exist.
By default, the server will use the sentence-transformers/all-MiniLM-L6-v2 embedding model to encode memories.
For the time being, only FastEmbed models are supported.
This MCP server can be used with any MCP-compatible client. For example, you can use it with
Cursor and VS Code, which provide built-in support for the Model Context
Protocol.
You can configure this MCP server to work as a code search tool for Cursor or Windsurf by customizing the tool
descriptions:
QDRANT_URL="http://localhost:6333" \
COLLECTION_NAME="code-snippets" \
TOOL_STORE_DESCRIPTION="Store reusable code snippets for later retrieval. \
The 'information' parameter should contain a natural language description of what the code does, \
while the actual code should be included in the 'metadata' parameter as a 'code' property. \
The value of 'metadata' is a Python dictionary with strings as keys. \
Use this whenever you generate some code snippet." \
TOOL_FIND_DESCRIPTION="Search for relevant code snippets based on natural language descriptions. \
The 'query' parameter should describe what you're looking for, \
and the tool will return the most relevant code snippets. \
Use this when you need to find existing code snippets for reuse or reference." \
uvx mcp-server-qdrant --transport sse # Enable SSE transport
In Cursor/Windsurf, you can then configure the MCP server in your settings by pointing to this running server using
SSE transport protocol. The description on how to add an MCP server to Cursor can be found in the Cursor
documentation. If you are
running Cursor/Windsurf locally, you can use the following URL:
http://localhost:8000/sse
[!TIP]
We suggest SSE transport as a preferred way to connect Cursor/Windsurf to the MCP server, as it can support remote
connections. That makes it easy to share the server with your team or use it in a cloud environment.
This configuration transforms the Qdrant MCP server into a specialized code search tool that can:
You can populate the database by storing natural language descriptions of code snippets (in the information parameter)
along with the actual code (in the metadata.code property), and then search for them using natural language queries
that describe what you're looking for.
[!NOTE]
The tool descriptions provided above are examples and may need to be customized for your specific use case. Consider
adjusting the descriptions to better match your team's workflow and the specific types of code snippets you want to
store and retrieve.
If you have successfully installed the mcp-server-qdrant, but still can't get it to work with Cursor, please
consider creating the Cursor rules so the MCP tools are always used when
the agent produces a new code snippet. You can restrict the rules to only work for certain file types, to avoid using
the MCP server for the documentation or other types of content.
You can enhance Claude Code's capabilities by connecting it to this MCP server, enabling semantic search over your
existing codebase.
Add the MCP server to Claude Code:
# Add mcp-server-qdrant configured for code search
claude mcp add code-search \
-e QDRANT_URL="http://localhost:6333" \
-e COLLECTION_NAME="code-repository" \
-e EMBEDDING_MODEL="sentence-transformers/all-MiniLM-L6-v2" \
-e TOOL_STORE_DESCRIPTION="Store code snippets with descriptions. The 'information' parameter should contain a natural language description of what the code does, while the actual code should be included in the 'metadata' parameter as a 'code' property." \
-e TOOL_FIND_DESCRIPTION="Search for relevant code snippets using natural language. The 'query' parameter should describe the functionality you're looking for." \
-- uvx mcp-server-qdrant
Verify the server was added:
claude mcp list
Tool descriptions, specified in TOOL_STORE_DESCRIPTION and TOOL_FIND_DESCRIPTION, guide Claude Code on how to use
the MCP server. The ones provided above are examples and may need to be customized for your specific use case. However,
Claude Code should be already able to:
qdrant-store tool to store code snippets with descriptions.qdrant-find tool to search for relevant code snippets using natural language.The MCP server can be run in development mode using the mcp dev command. This will start the server and open the MCP
inspector in your browser.
COLLECTION_NAME=mcp-dev fastmcp dev src/mcp_server_qdrant/server.py
For one-click installation, click one of the install buttons below:
Add the following JSON block to your User Settings (JSON) file in VS Code. You can do this by pressing Ctrl + Shift + P and typing Preferences: Open User Settings (JSON).
{
"mcp": {
"inputs": [
{
"type": "promptString",
"id": "qdrantUrl",
"description": "Qdrant URL"
},
{
"type": "promptString",
"id": "qdrantApiKey",
"description": "Qdrant API Key",
"password": true
},
{
"type": "promptString",
"id": "collectionName",
"description": "Collection Name"
}
],
"servers": {
"qdrant": {
"command": "uvx",
"args": ["mcp-server-qdrant"],
"env": {
"QDRANT_URL": "${input:qdrantUrl}",
"QDRANT_API_KEY": "${input:qdrantApiKey}",
"COLLECTION_NAME": "${input:collectionName}"
}
}
}
}
}
Or if you prefer using Docker, add this configuration instead:
{
"mcp": {
"inputs": [
{
"type": "promptString",
"id": "qdrantUrl",
"description": "Qdrant URL"
},
{
"type": "promptString",
"id": "qdrantApiKey",
"description": "Qdrant API Key",
"password": true
},
{
"type": "promptString",
"id": "collectionName",
"description": "Collection Name"
}
],
"servers": {
"qdrant": {
"command": "docker",
"args": [
"run",
"-p", "8000:8000",
"-i",
"--rm",
"-e", "QDRANT_URL",
"-e", "QDRANT_API_KEY",
"-e", "COLLECTION_NAME",
"mcp-server-qdrant"
],
"env": {
"QDRANT_URL": "${input:qdrantUrl}",
"QDRANT_API_KEY": "${input:qdrantApiKey}",
"COLLECTION_NAME": "${input:collectionName}"
}
}
}
}
}
Alternatively, you can create a .vscode/mcp.json file in your workspace with the following content:
{
"inputs": [
{
"type": "promptString",
"id": "qdrantUrl",
"description": "Qdrant URL"
},
{
"type": "promptString",
"id": "qdrantApiKey",
"description": "Qdrant API Key",
"password": true
},
{
"type": "promptString",
"id": "collectionName",
"description": "Collection Name"
}
],
"servers": {
"qdrant": {
"command": "uvx",
"args": ["mcp-server-qdrant"],
"env": {
"QDRANT_URL": "${input:qdrantUrl}",
"QDRANT_API_KEY": "${input:qdrantApiKey}",
"COLLECTION_NAME": "${input:collectionName}"
}
}
}
}
For workspace configuration with Docker, use this in .vscode/mcp.json:
{
"inputs": [
{
"type": "promptString",
"id": "qdrantUrl",
"description": "Qdrant URL"
},
{
"type": "promptString",
"id": "qdrantApiKey",
"description": "Qdrant API Key",
"password": true
},
{
"type": "promptString",
"id": "collectionName",
"description": "Collection Name"
}
],
"servers": {
"qdrant": {
"command": "docker",
"args": [
"run",
"-p", "8000:8000",
"-i",
"--rm",
"-e", "QDRANT_URL",
"-e", "QDRANT_API_KEY",
"-e", "COLLECTION_NAME",
"mcp-server-qdrant"
],
"env": {
"QDRANT_URL": "${input:qdrantUrl}",
"QDRANT_API_KEY": "${input:qdrantApiKey}",
"COLLECTION_NAME": "${input:collectionName}"
}
}
}
}
If you have suggestions for how mcp-server-qdrant could be improved, or want to report a bug, open an issue!
We'd love all and any contributions.
mcp-server-qdrant locallyThe MCP inspector is a developer tool for testing and debugging MCP
servers. It runs both a client UI (default port 5173) and an MCP proxy server (default port 3000). Open the client UI in
your browser to use the inspector.
QDRANT_URL=":memory:" COLLECTION_NAME="test" \
fastmcp dev src/mcp_server_qdrant/server.py
Once started, open your browser to http://localhost:5173 to access the inspector interface.
This MCP server is licensed under the Apache License 2.0. This means you are free to use, modify, and distribute the
software, subject to the terms and conditions of the Apache License 2.0. For more details, please see the LICENSE file
in the project repository.