Skip to main content

AgentGraph Feed

The AgentGraph Feed package provides both the AgentGraph connector and the shared HTTP server used to exchange graph mutation events.

Installation

Install the connector into the same Python environment as AgentGraph:

uv pip install agentgraph-connector-feed

The 0.1.x and 0.2.x connector releases support agentgraph-server>=0.6.1,<0.7. The 0.3.x releases require agentgraph-server>=0.7.0,<0.8. The 0.4.x and 0.5.x releases require agentgraph-server>=0.8,<0.9. The 0.6.x releases require agentgraph-server>=0.9,<1.0. The 0.5.x release line replaces agentgraph connector feed configure with enable and disable.

Configuration

Enable the connector with the feed server base URL, then inspect the saved configuration:

agentgraph connector feed enable http://localhost:8767
agentgraph connector feed status

Turn the feed off with disable, which keeps the saved URL, origin ID, and upsert setting so enable needs no arguments to turn it back on:

agentgraph connector feed disable
agentgraph connector feed enable

enable preserves the existing feed cursor, so events published while the feed was disabled are imported on the next poll. Pass agentgraph connector feed enable --restart to clear the cursor and start at the current feed tail instead.

The configured URL must not include /events or /events-feed. The connector adds /events and /events/tail when it publishes and polls respectively. For a deployed server behind a reverse proxy, configure the URL at which those endpoints are exposed.

The connector publishes new local observations, bookmark changes, and explicit deletions as best-effort HTTP requests. It polls the feed once per minute. Its first poll starts at the current feed tail, so existing events are not imported.

Entity upserts are ignored by default. Pass --publish-upserts to publish full committed entity and edge snapshots. Later enable runs inherit the stored setting; pass --no-publish-upserts to turn upsert delivery back off.

Server

Run the feed server locally with:

uv run agentgraph-feed-server

The server listens on port 8767 by default. Set AGENTGRAPH_FEED_SERVER_PORT to use another port. The event database defaults to <AGENTGRAPH_CONFIG_DIR>/agentgraph-feed-events.db, where the config directory defaults to ~/.agentgraph. Set AGENTGRAPH_FEED_DB to override the complete database path.

The server exposes GET /healthcheck, POST /events, GET /events, GET /events/tail, and DELETE /events?through=<sequence>. The delete endpoint purges only events at or before the supplied sequence, preserving events appended while a consumer processes its current page. It is intended for clients using the server as a one-producer, one-consumer queue.

Security

The feed is unauthenticated and should only be exposed on a trusted network.

Development

uv sync --locked
uv run pytest
uv run pyright
uv run ruff check .
uv run ruff format --check .

Metadata

Release files for agentgraph-connector-feed 0.6.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for agentgraph-connector-feed 0.6.0
File Size Uploaded Provenance Actions
agentgraph_connector_feed-0.6.0.tar.gz 15.7 kB With provenance Details for agentgraph_connector_feed-0.6.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for agentgraph-connector-feed 0.6.0
File Interpreter ABI Platform Actions
agentgraph_connector_feed-0.6.0-py3-none-any.whl Python 3 none any Details for agentgraph_connector_feed-0.6.0-py3-none-any.whl

Total release size: 26.9 kB

Release files / agentgraph_connector_feed-0.6.0.tar.gz

Download URL agentgraph_connector_feed-0.6.0.tar.gz
Size 15.7 kB
Tags Source
SHA-256 checksum
How to use checksums
52f9b6259c1b29e61c0673396434c9e655a6efa71bd1c0b3f8fca0ff4c9547d2
BLAKE2b-256 checksum
How to use checksums
32592bfb0ca750cd023f5c1b92b587c0019fe950d1d856e7f8138d1bb47063cd
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 19, 2026.

Transparency log

Release files / agentgraph_connector_feed-0.6.0-py3-none-any.whl

Download URL agentgraph_connector_feed-0.6.0-py3-none-any.whl
Size 11.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
4003451cadfe721f8574980a306f0b93ed054983d58de6cee531f6d5d9a2a39f
BLAKE2b-256 checksum
How to use checksums
37b2819f3b12caf5aa014ef4bc94be208084f8020db3d1e936cef17ef5521c7a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 19, 2026.

Transparency log

Security for agentgraph-connector-feed 0.6.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

All 2 files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for agentgraph_connector_feed-0.6.0.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 19, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page