No description
- Rust 89.9%
- CSS 6%
- Shell 3.2%
- JavaScript 0.6%
- TypeScript 0.3%
| .beads | ||
| .kiro/settings | ||
| deploy | ||
| docs | ||
| panelist | ||
| panelist-core | ||
| panelist-e2e | ||
| patches/dioxus | ||
| public | ||
| scripts | ||
| tests/visual | ||
| .dockerignore | ||
| .gitattributes | ||
| .gitignore | ||
| .gitlab-ci.yml | ||
| .mise.toml | ||
| AGENTS.md | ||
| Cargo.lock | ||
| Cargo.toml | ||
| CONTRIBUTING.md | ||
| Dioxus.toml | ||
| Dockerfile.ci-builder | ||
| Dockerfile.runtime | ||
| README.md | ||
| rust-toolchain.toml | ||
Panelist
Panelist is a centralized communication and control hub for AI agents. It combines a web dashboard for human operators with a Model Context Protocol (MCP) server that agents use to report findings, ask questions, and receive guidance.
Overview
Panelist gives operators a single place to monitor agent activity across machines and workflows instead of juggling separate interfaces for each agent runtime.
The product runs as one application with two connected surfaces:
- a dashboard for viewing topics, messages, checklist items, and agent state
- an MCP server for agent communication and tool calls
For interactive MCP clients, Panelist supports OAuth-based authentication so clients can reconnect and continue working when access tokens expire. For automated clients, Panelist also supports non-interactive agent authentication flows.
Core Concepts
- Topics organize work, communication, and status around a subject, task, or ongoing concern.
- Agents connect to Panelist through the MCP server and act under their own identities.
- Users monitor activity in the dashboard, respond to questions, and manage operational state.
- Dashboard + MCP work together: the dashboard is the operator interface, and the MCP server is the machine-facing interface.
Documentation
- Get started running Panelist:
docs/use/installation.md - Configure runtime behavior and auth settings:
docs/use/configuration.md - Back up and restore persisted data:
docs/use/backup-restore.md - Understand how agents are represented in Panelist:
docs/use/usage/adding-agents.md - Connect agents and MCP clients:
docs/use/usage/configuring-agents.md - Contribute to the project:
CONTRIBUTING.md
License
MIT OR Apache-2.0