Pixeltable Cloud is in Limited Beta. Email [email protected] if you are interested.

The unified multimodal backend for AI data apps in one Python file

Quickstart | Documentation | CLI | Cloud | Skill | get-started.md | skill.md | llms-full.txt | Discord

pip install 'pixeltable[serve]'
pxt init
pxt service example --out app.py
pxt schema update app.py my_app
pxt service update app.py my_app

import pixeltable as pxt
import pixeltable.functions as pxtf
from pixeltable.serving import FastAPIRouter
TableModel = pxt.model_base()
@pxt.udf                                        # a Python function the columns below can call
def excerpt(text: str, n: int = 12) -> str:
    return text if len(text) <= n else f'{text[:n]}...'
class Docs(TableModel, name='docs'):
    doc_id = pxt.Column(value=pxtf.uuid.uuid7(), primary_key=True)  # a generated key: provided automatically on insert
    title: pxt.String
    body: pxt.String | None
    title_upper = pxtf.string.upper(title)      # an assignment: computed on insert and on update
    summary = excerpt(title)                    # a computed column over a udf this file defines
ingest = FastAPIRouter(name='ingest')
ingest.add_insert_route(                        # POST /docs inserts and returns the computed columns
    Docs, path='/docs', inputs=[Docs.title, Docs.body], outputs=[Docs.doc_id, Docs.title_upper, Docs.summary]
)
ingest.add_compute_route(Docs, path='/titles', inputs=[Docs.title], outputs=[Docs.title_upper])

The same file holds pxt.Image, pxt.Video, pxt.Audio, or pxt.Document columns, and a computed column over one of them is another assignment: media pipelines, RAG. The port is assigned, so read it back rather than hardcoding it:

URL=$(pxt service list --json | jq -r '.[0].endpoint')
curl -X POST "$URL/docs" \
  -H 'Content-Type: application/json' \
  -d '{"title": "Hello", "body": "world"}'
{"doc_id":"...","title_upper":"HELLO","summary":"Hello"}

The same file runs on Pixeltable Cloud. Create an API key in the Cloud dashboard, set PIXELTABLE_API_KEY, name the database in pixeltable.toml, then target it by URI. pxt db update creates or updates the hosted database; it does not insert rows. pxt service run is local only and cannot target Cloud.

pxt db update pxt://org:mydb
pxt schema update app.py pxt://org:mydb
pxt service update app.py pxt://org:mydb

A @pxt.udf in that same app.py is in the image pxt db update builds.

Chat agent or video search

uvx pixeltable-new copies one app from the starter kit. The default copy is a chat app; pass agent as the last argument to pxt schema update. --video copies video search; pass videointel.

uvx pixeltable-new myapp
cd myapp
uv sync
pxt schema update app.py agent
pxt service update app.py agent

Inserting into the knowledge table needs no API key. The /ask route needs ANTHROPIC_API_KEY.

To mount the routes on an existing FastAPI app, app.include_router(...). HTTP serving. To skip endpoints, run pxt schema update, insert from Python, then export_sql. Self-hosting.

Coding agents

Hand the agent get-started.md. That playbook installs the package, the skill, and MCP. How the skill writes app.py: AI coding agents. Docs capability file: skill.md.

npx skills add pixeltable/pixeltable-skill

The skill writes a TableModel in app.py. If the agent writes create_table in application code, names the file schema.py, or writes the removed command pxt serve, the installed skill is stale: reinstall npx skills add pixeltable/pixeltable-skill.

Notebooks and tests still use pxt.create_table(). An app puts tables in app.py and creates them with pxt schema update.

License

Apache 2.0. Contributing · Discord