Vote for Postgres extensions

[kiwicopple]

This issue is for tracking requests/demand for Postgres Extensions. Give a thumbs up if you want an extension added.

Please refrain from adding comments unless you are adding an extension which isn't already listed.

We will be sorting by Top here, so please upvote the extension, the emojis don't count as upvotes. Don't create duplicate requests, so that the votes aren't split.

[kiwicopple]

plv8 - https://github.com/plv8/plv8

[kiwicopple]

released 🚢

[ozum]

@kiwicopple, could you please hide messages when a requested extension is added, so we can easily see the remaining ones?

[skamensky]

Removed 😞

[kiwicopple]

PL Python

edit: this cannot be supported on Supabase because it is an untrusted extension. See here for more details
https://github.com/orgs/supabase/discussions/2239

[skwaugh]

Was support for this recently dropped? Can't find it in a new pg13 supabase project...which means it is not enabled during initial build. Is this temporary or the norm going forward?

[chrichrichri]

Any news on the plpython extension we would really love to use this in our project.

[burggraf]

@inian Any word on this python extension?

[inian]

#2239 (comment)

[shah]

PL Java would be great - https://github.com/tada/pljava.

Also, ARG based build which would enable/disable certain plugins might also be helpful.

Great work by the way, love the Dockerfile.

[kenahrens]

Timescale - https://github.com/timescale/timescaledb

[georgewritescode]

Following up to also note that timescaledb_toolkit would be very helpful given it has many of the functions required to use timscaledb. Also referenced here: https://github.com/orgs/supabase/discussions/17108

[encima]

Timescale-toolkit is under the TSL License and cannot be offered by Supabase (or any Database-as-a-service)

[riderx]

Timescale is useless in supabase environments…
That why I made a PR to have a columnar db alternative:
https://github.com/orgs/supabase/discussions/679#discussioncomment-7791898
please have a look and if that fit your use case upvote it

[bitnom]

IMO some alternative to timescale should have been used. The licensing is bad, and the version you can use in supabase cloud is not good.

[riderx]

The official alternative has been announced today: https://supabase.com/blog/analytics-buckets

[shah]

plpgsql_check for linting and validating SQL code quality.

[kiwicopple]

Added 👍

[felixyz]

wal2json

[dragarcia]

Added via Supabase Postgres v13.3.0.

[Takashiidobe]

I would like to see pg_cron on here, although I don't know about the licenses.

[kiwicopple]

Added 👍

[psolymos]

PL/R would be great! https://github.com/postgres-plr/plr and https://access.crunchydata.com/documentation/plr/8.3.0.18/

[Vadorequest]

https://github.com/furstenheim/is_jsonb_valid seems useful for enforcing JSON schemas in columns using the JSON type.

It's not in the Database > Extensions yet, I suggest it is added.

It seems to be more performant than its https://github.com/gavinwahl/postgres-json-schema counterpart. (from the author)

[chipilov]

FYI, since the https://github.com/gavinwahl/postgres-json-schema extension is implemented in PL/pgSQL, you can enable it manually in your instance by running its script against your db: https://github.com/gavinwahl/postgres-json-schema/blob/master/postgres-json-schema--0.1.1.sql (you just need to remove/replace the references to @extschema@, since that's only available for extensions).

[cananda]

https://github.com/supabase/pg_jsonschema :)

[quantpioneer]

pgvectorscale

[andreasfranzen]

I'm also in need of this as pgvector's hnsw is problematic with filters.

[krystian-dev548]

keccak
Useful for saving hash on blockchain

[AlexandreSi]

pg_rational
Useful to store position of a row in an array

[Lqm1]

pg_bigm
Easier to handle, like pg_trgm, and better for full text searching in Japanese.
PGroonga is too complex. If you use Prisma ORM together, it won't work properly.

[Wobbel26]

I could really use mysql_fdw for multiple migrations trajects.

https://github.com/EnterpriseDB/mysql_fdw

[wenerme]

seems supabase not intended to introduce the mysql_fdw, I made my own by forking the https://github.com/wenerme/wrappers
My prebuild image https://github.com/wenerme/dockerfiles/tree/master/postgres/postgres is wener/postgres

[DerekBBell]

Ability to use PostgreSQL 18 UUIDv7, both to set as the default UUID method and implemented as a function - uuid_generate_v7().

[encima]

You can use this pure sql extension from database.dev to get support already :)
https://database.dev/cem/uuidv7 (even if you do not use the tool, you can just execute the SQL against your DB)

[peekknuf]

We got analytics in the ability to connect to DuckDB

But the beautiful thing is that you can already have it within your regular Postgres with pg_duckdb
Having to connect to an external engine is tedious, while having just SET duckdb.force_execution = true; to run your 5 CTEs 7 joins analytical query is just so nice.
I've ran the benchmarks and it's pretty much equivalent to the speed of the binary, but even if it were just half as performant the convenience of one liner to run analytics is unmatched.

Now with the release of pg_lake (it's shiny but almost too new, i've yet to test it, but on paper sounds fantastic) there is an opportunity to make Supabase "the everything Postgres" platform.

This in particular sounds almost too good

pg_lake introduces a new Iceberg table type where Postgres itself acts as the catalog.

But at this point i'm curious how it would work with Analytical Buckets that are being developed right now

[alfalcon90]

pg_semver
This library contains a single PostgreSQL extension, a data type called "semver".

[maxios]

timescaledb (latest version supported pg:17)

[magick93]

I think timescale now supports pg18

[encima]

I am afraid Timescale support is being deprecated from 17 onwards, we recommend to use pg_partman

[Soviut]

Add reoccurring rule (rrule) support with https://github.com/volkanunsal/postgres-rrule

RRules are very useful for representing things like calendar entries that repeat (eg: Standup Meeting every Monday at 10am) without needing to materialize all those repeated entries. They also support limiting the number of repetitions by count or end date.

Rules are plain text key-value pairs separated by a ; with each key and value being separated by an =.

FREQ=WEEKLY;BYDAY=MO;COUNT=4

The syntax can be explored with this rules generator.

https://www.addevent.com/c/documentation/tools/recurring-rule-generator

While it is possible to perform rrule calculations at the application level without this extension, having it installed means it's much easier to select a date range that occurs long after the initial entry and know that you'll capture repeated entries.

[encima]

It does not solve your problem directly but Postgres is actually quite good at handling overlapping time periods docs

[giorgiogross]

BM25 https://github.com/timescale/pg_textsearch

[SteveLauC]

bump!

[lei334b]

+1 This would be incredibly helpful for building knowledge bases.

[jpargudo]

"The Cryptic Elephant", for Transparent Column Encryption: https://gitlab.com/dalibo/the_cryptic_elephant at Dalibo labs (PostgreSQL License).

Ongoing project, but very promising. Will comment this post when they will go "production ready".

[allpwrfulroot]

The new pg_clickhouse extension, plz! Analytics directly from Postgres (views/RPCs) without SQL rewrites and reducing data-pipeline complexity.

[encima]

We have support for Clickhouse using our wrappers extension!

[csenio]

pg_roaringbitmap

[leotorres6771-afk]

As a workaround u can try disabling this option

…

On Tue, Mar 17, 2026, 4:37 AM Chris Gwilliams ***@***.***> wrote: We have support for Clickhouse <https://fdw.dev/catalog/clickhouse/> using our wrappers extension! — Reply to this email directly, view it on GitHub <#679 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/B72L6ATCUBZKOPXIZDEUANT4REFEXAVCNFSM4XHHAAGKU5DIOJSWCZC7NNSXTOSENFZWG5LTONUW63SDN5WW2ZLOOQ5TCNRRG42DAOJT> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>