Built for complex, long-running research
Start with the module you need, and keep the scope, experiment results, and delivery evidence traceable.
Project-level research agent
Keep independent ModuleRuns, experiment records, and artifacts in one project. Continue from a selected candidate while preserving earlier branches.
20+ LLM providers, out of the box
OpenAI, Anthropic, OpenRouter, DeepSeek, Azure, Ollama, vLLM and more. Configure ordered model candidates and team role providers; fallback depends on actual runtime support.
CLI / Web / Desktop — three form factors
Run experiments from the command line, watch progress in a browser, or launch the Electron app. They share one config and sync state in real time.
13 channel integrations
Lark, Slack, Telegram, DingTalk, email, Matrix and more. @-mention the agent in a group to run experiments, ship reports, and reply with results — no browser needed.
Execution approval and experiment control
Approve the scope before execution. In manual mode, review each experiment before continuing or summarizing. Modules validate inputs and outputs; legacy plans retain field checks.
Local-first, self-host friendly
Keep projects in their selected directories. Use a local service, Docker, or a remote gateway, with desktop controls for the local engine.
A workbench built for agents
MiraUI combines chats and projects, independent module runs, experiment records, and delivery in one workspace. Choose a direction, approve the scope, compare experiments, and turn selected evidence into a report, paper, or presentation.
- Real-time sync: WebSocket events and REST snapshots keep module and experiment state in sync.
- Same code for Web & Desktop: develop in the browser and use Electron for desktop service controls.
- Approve and compare: approve the scope, review experiments, and preserve results from different branches.
