Slack-native agent routing OSS

Bring AI agents into the Slack conversations where work already happens.

Agents party is an OSS foundation inspired by Party on Slack. It helps teammates, AI, and operators collaborate in the same Slack workflow across summaries, search, translation, file understanding, and specialist tools.

Slack Bolt Multi-provider LLM routing PostgreSQL Terraform

From Party on Slack to OSS

Agents party extracts the Party on Slack idea of bringing AI into Slack conversations into an OSS foundation that operators can run in their own workspace and infrastructure.

Inherited idea

Collaboration inside Slack, not another app

AI work should not disappear into a private screen. The foundation keeps teammates and agents in the same channel and thread context.

OSS foundation

An operator-owned implementation base

Agents party is not the commercial Party on Slack app. It is the open-source base for teams that want to extend and operate Slack-native AI collaboration themselves.

What you can build in Slack

Agents party gives teams a foundation for combining conversation, search, translation, file understanding, and specialist tools from Slack mentions and threads.

Agent routing

Ask agents in channels and threads

Resolve agents and models from app mentions, assistant threads, and active thread follow-up context.

Workspace tools

Bring search, translation, and Canvas back to work

Extend Slack search, reaction translation, Canvas output, RSS summaries, and other tool paths without leaving the workspace flow.

Specialists

Handle files and business workflows

Call specialist tools for images, audio, PDFs, Salesforce PDF workflows, and other task-specific work from the agent runtime.

Own your Slack AI stack

The value of the OSS foundation is that operators can design the boundaries around their Slack workspace, provider keys, model routing, deployment, and data handling.

Self-host

Run it on infrastructure you choose

Build deployment and operations around the Node.js runtime, PostgreSQL, Redis-backed workers, and Terraform for your environment.

Credentials

Control provider keys and model costs by workspace

Separate Slack OAuth, workspace credentials, and multi-provider routing so teams can choose providers and models by use case and workspace policy.

Data boundary

Review data handling assumptions before operating

Use Data processing and Security to evaluate the services and provider terms you enable.

Open source signals you can inspect

Before adopting the project, operators can review the license, security notes, data processing guidance, contribution path, and issue tracker on GitHub. Agents party is a pre-1.0 OSS application under the MIT License.

Project

Status and contribution

Track work through Issues and use Contributing to understand how to participate.

Product philosophy

AI should not be trapped in a separate app. Agents party turns the Party on Slack philosophy into an operator-owned OSS foundation for Slack workspaces.

Conversation

Teammates, AI, and you

Agents participate in Slack channels and threads, not only private one-to-one chats.

Routing

No single-model assumption

Route across OpenAI, Anthropic, Google, Vertex AI, and other providers by use case and workspace policy.

Operator control

Clear runtime ownership

Slack OAuth, workspace credentials, PostgreSQL, and Terraform boundaries are explicit from the start.

Agents party is a pre-1.0 OSS application. It is not the commercial Party on Slack app; it is an implementation foundation inspired by the same Slack-native collaboration philosophy.