Extensions

DB9 ships all extensions compiled into the server binary. There is no runtime dynamic loading, which is what buys you the guarantees below: there is nothing to install, no build step, no version skew between two databases, and every extension is available the moment a database exists. CREATE EXTENSION records install metadata per database rather than loading anything.

The trade is that the set is fixed. CREATE EXTENSION accepts exactly the names on this page and raises 42704 extension is not available for every other name — PostGIS, pg_trgm, ltree, citext and the rest of the PostgreSQL contrib ecosystem included.

CREATE EXTENSION accepts eleven names, and they fall into two groups:

• Eight functional extensions — fs9, parquet, vector, embedding, zhparser, http, pg_cron and uuid-ossp (uuid_generate_v4() only). These add types, functions, or operators.
• Three metadata-only shims — hstore, pgcrypto and plpgsql. CREATE EXTENSION succeeds and writes a pg_extension row so dumps and ORM bootstrap scripts restore unchanged, but none of them registers a function of its own, and the hstore type name that hstore adds to pg_type cannot hold a value. Read the pgcrypto caution before you depend on any of them.

This page is the authoritative list; other pages link here rather than restating it.

Extensions by Task

Files and Data

Extension	What it does	Docs
fs9	Query, read, and write the database filesystem from SQL	fs9 reference
parquet	Read Parquet files from the filesystem or HTTP URLs into SQL tables	Parquet import

Search

Extension	What it does	Docs
vector	pgvector-compatible similarity search with HNSW indexing	Vector Search
embedding	Server-side text embedding via EMBEDDING() and EMBED_TEXT() — no external API needed	Vector Search
zhparser	Chinese full-text search tokenizer (jieba-rs, always available)	Full-Text Search

English and general-purpose full-text search is not an extension — tsvector, tsquery, to_tsvector(), websearch_to_tsquery(), GIN indexes on tsvector and the rest of the FTS surface are core server features, so there is no fts name to pass to CREATE EXTENSION. See Full-Text Search. zhparser is the one FTS-related extension descriptor, and it covers Chinese tokenization only.

Integration and Automation

Extension	What it does	Docs
http	Make HTTP requests from SQL (GET, POST, PUT, DELETE, HEAD, PATCH as table functions)	HTTP client
pg_cron	Schedule SQL statements to run on a cron expression	pg_cron

Compatibility

Extension	What it does	Docs
uuid-ossp	uuid_generate_v4() only — the other uuid-ossp functions (uuid_generate_v1(), uuid_generate_v3(), uuid_generate_v5(), uuid_nil(), …) fail with 42883	uuid-ossp
hstore	Metadata only — hstore values are not supported (0A000). Use JSONB	hstore
pgcrypto	Metadata only — registers no functions. CREATE EXTENSION pgcrypto succeeds so Supabase and ORM bootstrap scripts restore unchanged, but crypt(), gen_salt(), hmac(), encrypt()/decrypt(), pgp_sym_encrypt() and gen_random_bytes() are not available	caution below
plpgsql	Metadata only — PL/pgSQL is compiled into the server and works with or without CREATE EXTENSION plpgsql; the statement exists so restored dumps do not fail	caution below

Enabling Extensions

Most extensions require CREATE EXTENSION before use. A few are pre-enabled or always available.

Pre-enabled (available on every new database)

These extensions are installed automatically when a database is created:

SQL

-- Already enabled — no CREATE EXTENSION needed
SELECT http_get('https://httpbin.org/get');
SELECT count(*) FROM cron.job;

Note

http_get and the other HTTP functions return JSONB — call them as scalar functions (SELECT http_get(...)), not as table-valued functions. See HTTP client for the full reference.

pg_cron job execution is unavailable in the current release

The cron schema functions are installed and callable immediately, but cron.schedule(...) currently returns an error — job scheduling/execution requires the worker subsystem, which is disabled for this release. See pg_cron for details.

• http — HTTP client functions available immediately
• pg_cron — Extension pre-installed; catalog and management functions (cron.job, cron.unschedule, etc.) are available immediately, but job scheduling/execution is unavailable in the current release

Require CREATE EXTENSION

These extensions must be explicitly enabled — for fs9, only its table functions depend on it:

SQL

CREATE EXTENSION IF NOT EXISTS fs9;
CREATE EXTENSION IF NOT EXISTS embedding;
CREATE EXTENSION IF NOT EXISTS parquet;

▶ Run

• embedding and parquet — until enabled, calls fail with ERROR: extension "embedding" is not installed. Run: CREATE EXTENSION embedding (0A000), and the same message naming parquet for extensions.read_parquet(...).
• fs9 — the scalar functions (fs9_write, fs9_read, fs9_append, fs9_mkdir, fs9_exists, fs9_remove and the other fs9_* file functions) work on a new database without CREATE EXTENSION, bare or extensions.-qualified, and calling them does not add fs9 to pg_extension. The table functions — extensions.fs9(), extensions.fs9_events(), extensions.fs9_jg(), extensions.fs9_storage_stats() and extensions.fs9_cached_storage_stats() — fail with ERROR: unsupported table-valued function: fs9 (0A000, naming the function you called) until you run CREATE EXTENSION fs9.

Metadata shims (built-in, CREATE EXTENSION optional)

These need no CREATE EXTENSION — what they provide is built in, or, for hstore, not implemented at all — but running it records metadata in pg_extension for ORM and tool compatibility:

Run CREATE EXTENSION for ORM compatibility

Even though vector and uuid-ossp work without CREATE EXTENSION, running it registers the extension in pg_extension. ORMs like Prisma and Drizzle check this table to detect capabilities — running CREATE EXTENSION prevents false-negative compatibility errors. The same applies to hstore, pgcrypto and plpgsql, which exist only as catalog entries — but register hstore only if a tool needs it: while it is registered, pg_dump cannot dump the database.

• vector — The vector type and operators (<->, <#>, <=>) work as built-ins. HNSW indexes can be created and the planner uses them for similarity search when the probe is a literal or a bound parameter and the query has a LIMIT and no WHERE clause — see Vector Search. CREATE EXTENSION vector is recommended so ORMs that check pg_extension see it as installed.
• uuid-ossp — uuid_generate_v4() is built in, and it is the only uuid-ossp function DB9 has: uuid_generate_v1(), uuid_generate_v3(), uuid_generate_v5() and uuid_nil() fail with 42883. CREATE EXTENSION "uuid-ossp" records metadata for pg_dump/pg_restore compatibility.
• hstore — Extension metadata only: hstore values are not supported, so any use of the type fails with 0A000 type "hstore" is not supported and the hint Use jsonb instead. — store key-value data in JSONB. While the extension is registered, pg_dump fails with column "typanalyze" does not exist; DROP EXTENSION hstore clears it.
• pgcrypto — Metadata only, so Supabase and ORM bootstrap scripts that run CREATE EXTENSION pgcrypto succeed unchanged. It registers no functions of its own: gen_random_uuid() and digest() are DB9 built-ins that work with or without it. The rest of pgcrypto is not implemented — see the caution below.
• plpgsql — PL/pgSQL is compiled in. CREATE EXTENSION plpgsql records metadata so restored dumps (which almost always contain it) do not fail.

CREATE EXTENSION pgcrypto succeeds, but the crypto functions do not exist

Running CREATE EXTENSION pgcrypto returns CREATE EXTENSION and adds a row to pg_extension, so a schema restore will not fail on it. It does not make pgcrypto’s functions available. crypt(), gen_salt(), hmac(), encrypt()/decrypt(), pgp_sym_encrypt() and gen_random_bytes() all raise 42883 function ... does not exist when called.

Only gen_random_uuid() and digest() work, and both are DB9 built-ins available on a database with no extensions installed. Hash a value with digest(data, 'sha256'), sha256() or md5(); move password hashing and symmetric encryption into your application layer.

Always available (no-op)

• zhparser — The Chinese full-text tokenizer is compiled in and always active. CREATE EXTENSION zhparser is accepted but has no effect.

Extension Schemas

Extensions install their functions into specific schemas:

Extension	Default schema	Access pattern
http	extensions	http_get(...) for the scalar JSONB form; SELECT * FROM extensions.http_get(...) for the table form. The qualifier is not interchangeable — see http
fs9	extensions	fs9_read(...) or extensions.fs9_read(...) — both resolve on the default search_path
embedding	extensions	embedding(...) — the bare name; extensions.embedding(...) does not resolve
parquet	extensions	SELECT * FROM extensions.read_parquet(...) — table-valued, FROM position only
pg_cron	cron	cron.schedule(...), cron.unschedule(...)
vector	public	Operators and type available directly
uuid-ossp	public	uuid_generate_v4() available directly
hstore	public	Metadata only — hstore values are not supported
pgcrypto	extensions	Metadata only — registers no functions (gen_random_uuid(), digest() are built-ins)
plpgsql	pg_catalog	Metadata only — PL/pgSQL is compiled in

For vector, uuid-ossp, hstore, pgcrypto and plpgsql, CREATE EXTENSION ... SCHEMA accepts only the schema in this table; naming another fails with 0A000 extension "vector" must be installed in schema "public". The vector type, its operator classes and its operators are built in, so write them unqualified (OPERATOR(public.<=>) fails with internal error): public.vector(3) and extensions.vector(3) both fail with 42704 type "public.vector" does not exist (or "extensions.vector"), and extensions.uuid_generate_v4() fails with 42883. pg_dump writes these names schema-qualified — rewrite them before an import, as described in Migrate from PostgreSQL and Migrate from Supabase.

Choosing the Right Extension

I want to…	Use
Store and search vector embeddings	vector + embedding
Build a RAG pipeline without external APIs	embedding + fs9
Call an external API from SQL	http
Query files on the database filesystem	fs9
Watch filesystem changes in real time	fs9 events + CLI watch
Import Parquet data	parquet
Schedule a recurring cleanup or report	pg_cron
Search text content	Full-Text Search
Generate UUIDs	uuid-ossp

Next Steps

• Architecture — How extensions fit into the DB9 server
• SQL Reference — SQL engine compatibility and features
• RAG with Built-in Embeddings — End-to-end tutorial using embedding() and vector search
• Analyze Agent Logs with fs9 — Tutorial: write and query agent logs from SQL
• HTTP from SQL — Tutorial: call APIs, webhooks, and enrich data from SQL
• Scheduled Jobs with pg_cron — Tutorial: periodic tasks, monitoring, and job management
• Agent Workflows — Extensions in agent pipelines
• CLI Reference — db9 db sql for running extension commands