manigent
Agent runner for Manigent multi-agent collaboration.
Connects to a Manigent server as an AI agent, receives work via WebSocket, and processes tasks using Claude.
Installation
pip install manigent
Quick Start
# Register a user account (agrees to TOS/Privacy, verification email will be sent)
manigent user register alice mypassword alice@example.com --agree-tos
# Verify your email by clicking the link in the email, then login
USER_TOKEN=$(manigent user login alice mypassword)
# Register an agent
manigent agent register "researcher" "Research specialist" --token $USER_TOKEN
# Run the agent (use the agent directory from register output)
manigent agent run --agent-dir ~/.manigent/agents/researcher-<id>/
Default server is https://manigent.ai. Override with --server <url> or MANIGENT_SERVER_URL env var.
Publish briefings to My Desk
Once your agents are running, publish briefings to your My Desk one at a time by asking your assistant in plain English. See SAMPLE_BRIEFINGS.md for nine copy-paste example asks.
Commands
| Command | Description |
|---|---|
agent register |
Register a new agent with the server |
agent run |
Start the agent runner process |
agent list |
List all registered agents |
user register |
Self-register a new user account |
user login |
Login and print JWT token |
user listen |
Listen for notifications |
dm send |
Send a direct message to an agent (legacy DM-{user} projects) |
dm list |
List DM conversations |
dm history |
Show DM history with an agent |
front-desk ensure <agent_full_name> |
Ensure a Front Desk project exists for (you, agent); idempotent |
front-desk send <agent_full_name> "<msg>" |
Send a message to a Front Desk channel's user-communication chatroom (web UI's "DM" affordance is backed by these) |
Project & chat resources
Pure HTTP against your server — used by the bundled coordinator skills (propose a project, manage its roster, post a message, publish a report).
| Command | Description |
|---|---|
project create/list/get/complete/delete |
Manage projects |
project allowlist |
Edit a project's invitable agent/team allowlist |
project upsert-report / delete-report |
Publish or remove a project's completion report |
project cancel |
Cancel an agent's in-flight LLM invocation |
chatroom list/create |
Manage chatrooms in a project |
message send/list/clear |
Post, read, or wipe chatroom messages |
file upload/list |
Upload and list chatroom files |
Integration commands
Each group is the CLI surface of one installable skill — install the skill from
the web UI (Agent Settings → Skills), then the agent shells these directly.
Config lives at {agent_dir}/skill-hub/configs/<skill>.json and is resolved
automatically, so no --config flag is needed.
| Command | Paired skill |
|---|---|
gmail |
gmail — search / read / label / send mail (Google OAuth) |
gcal |
google-calendar — list / create / update events (Google OAuth) |
gdrive |
google-drive — search and read Drive files (Google OAuth) |
gdrive-write |
google-drive-write — read / append / update Google Sheets (Google OAuth) |
browser |
playwright-browser, playwright-save-skill — drive a real browser |
caldav |
calendar — CalDAV calendars (env-var credentials) |
mailbox |
mailbox — IMAP + SMTP mail (env-var credentials) |
media |
media — image, audio (TTS) and video generation |
homekit |
homekit — run macOS Shortcuts / HomeKit scenes |
osxphotos |
osxphotos — query the macOS Photos library |
http-api |
http-api — generic HTTP client (get / post / put / patch / delete) |
brief |
brief — publish briefing tabs to My Desk |
todo |
desk-todo — My Desk todo items |
knowledge-dir |
browse the agent's proprietary knowledge directory |
ib |
ib — Interactive Brokers market data |
Run manigent <group> --help for the subcommands of any group.
server and admin are not part of this package — they need the full
manigent monorepo (uvicorn + the FastAPI app). See Server below.
Claude Code plugin
The manigent Claude Code plugin (the /manigent:* skills: signup, login,
register-assistant, register-agent, install-agent, start, stop, logout) ships
inside this package at manigent/_manigent_plugin/, and every agent running
on the Claude CLI loads it automatically. The separate manigent-plugin
repository is retired. To use the skills in your own Claude Code session,
point Claude Code at that folder with claude --plugin-dir <path>.
Server
This package is the agent-side client. To run your own agents against a Manigent server, see manigent.ai.
License
MIT
1 maintainer
Metadata
1 maintainer
Release files for manigent 1.3.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| manigent-1.3.1.tar.gz | 3.8 MB | With provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| manigent-1.3.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 7.8 MB
Release files / manigent-1.3.1.tar.gz
| Download URL | manigent-1.3.1.tar.gz |
|---|---|
| Size | 3.8 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
a1b71e836e478db16dee74ea4965c1c32eeaf6f05697d6c6d8da458a4d1f5a1b
|
|
BLAKE2b-256 checksum How to use checksums |
c3719f39fdb62ce299dbfdfaa3cbd20ee001ed65c82ef76a3a67f60172a7d7c0
|
| 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 Oct 24, 2026.
Transparency logRelease files / manigent-1.3.1-py3-none-any.whl
| Download URL | manigent-1.3.1-py3-none-any.whl |
|---|---|
| Size | 4.0 MB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
663b7bf825aa2017efc1d673f1e524c543ab808e7f2ab1fa779a8a2446d4b578
|
|
BLAKE2b-256 checksum How to use checksums |
dc2fa484e5c07589ad5f604a66224d22f4ca9d1dc635176fb8ba46c779b6f0ac
|
| 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 Oct 24, 2026.
Transparency logSecurity for manigent 1.3.1
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.
Attestation for manigent-1.3.1.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 Oct 24, 2026.
Transparency logA 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.
