Mattermost Agents
An open-source AI agent plugin for self-hosted Mattermost that connects to Ollama, vLLM or cloud providers to search and summarize team conversations.
Mattermost Agents brings AI assistants into a self-hosted Mattermost workspace, with a choice of local models or cloud providers. It's for teams that want help with long discussions and meeting recordings while controlling where their collaboration data goes. The plugin is open source under Apache 2.0.
Teams can run models on their own infrastructure through Ollama, vLLM or an OpenAI-compatible API. They can also connect to OpenAI, Anthropic, Azure or Amazon Bedrock. With a local backend, collaboration data can stay within the team's network; choosing a cloud provider sends requests to that service. The plugin needs network access to whichever model provider it uses.
Agents summarize threads and channels, extract action items, and transcribe and summarize meeting recordings. Natural language search finds relevant content across the Mattermost instance. People can chat directly with an agent or tag one in an existing conversation, and teams can create separate assistants with different roles and capabilities.
Its connection to Mattermost's workflows goes beyond chat. Agents can work with Playbooks and use MCP connectors to call internal systems, with user approval before tools run. Access controls govern who can use agents, and token accounting tracks model usage. A Go client library also lets other Mattermost plugins and server features use the plugin's LLM connection. Semantic search requires PostgreSQL with the pgvector extension.
