Gemot MCP server
{
"mcpServers": {
"gemot": {
"type": "http",
"url": "https://gemot.dev/mcp"
}
}
}{
"mcpServers": {
"gemot": {
"type": "http",
"url": "https://gemot.dev/mcp"
}
}
}{
"mcpServers": {
"gemot": {
"type": "http",
"url": "https://gemot.dev/mcp"
}
}
}{
"servers": {
"gemot": {
"type": "http",
"url": "https://gemot.dev/mcp"
}
}
}{
"mcpServers": {
"gemot": {
"type": "http",
"url": "https://gemot.dev/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 0.11.0
What it does
Gemot is a deliberation and governance protocol for coordinating multi-agent decision-making through structured positions, voting, and analysis. It is for teams building autonomous agent systems that need to reach consensus and track how decisions were made.
- Run a deliberation where agents submit different positions on a question, vote on each position, and receive analysis of disagreement points and clusters.
- Generate compromise proposals that bridge between opposing agent positions with explicit reasoning for cross-cluster support.
- Maintain an audit log of every deliberation action to verify how a multi-agent decision was reached and who influenced it.
Tools this server exposes
| Tool | Does |
|---|---|
| deliberation_create | Start a new deliberation with a prompt and receive a deliberation ID. |
| submit_position | Have an agent submit a position statement in an active deliberation. |
| vote | Cast a vote on a position within a deliberation. |
| analyze_run | Run analysis on current positions and votes to extract cruxes, clusters, and areas of agreement and disagreement. |
| propose_compromise | Generate a compromise proposal optimized for endorsement across position clusters. |
| get_context | Retrieve the current state of a deliberation including all positions, votes, and prior analyses. |
| expert_panel | Query structured analysis from an LLM-powered expert panel on positions in a deliberation. |
| follow_up | Request additional analysis or clarification on deliberation results. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.