council-mcp
Integrates with Google Gemini as an advocate LLM, querying Gemini via its OpenAI-compatible endpoint to answer deliberation questions in parallel with other models.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@council-mcp Ask the council whether we should migrate our database to Postgres 16 this weekend."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
council-mcp
A Model Context Protocol server that runs multi-model deliberation: several LLMs answer the same question in parallel, then a judge model synthesizes their perspectives into one answer.
It is built for the cases where a single model is not enough — architecture decisions, debugging strategy, high-risk operational changes, or any question where hallucination risk matters. It also treats "the model returned HTTP 200" as a claim to verify, not proof that the model actually said anything useful.
Parallel advocates — 2–4 models queried concurrently with a hard per-call wall-clock cap.
Judge synthesis — a separate model produces consensus, dissent, a final verdict, and a confidence level from the substantive advocate responses.
Degraded-response detection — catches empty bodies,
<think>/<thinking>reasoning wrappers, DeepSeek's end-of-thinking marker, and untagged chain-of-thought leaks that used to sail through as valid answers.Bounded retry and backfill — degraded or failed advocates get one retry; unused providers backfill remaining slots, all inside a wall-clock budget so a deliberation cannot run past the caller's deadline.
Cost and latency telemetry — every response reports provider, model, latency, estimated cost, finish reason, and degradation status.
How it works
question
│
├──► advocate A ──┐
├──► advocate B ──┤ parallel, hard-capped
├──► advocate C ──┤
└──► advocate D ──┘
│
strip reasoning wrappers
classify degraded output
│
┌──────────┴──────────┐
│ retry degraded once │
│ backfill from pool │ wall-clock budget
└──────────┬──────────┘
│
substantive responses only
│
judge model
│
synthesis + metadataIf no advocate produces substantive content after retry/backfill, the server
fails closed: it returns All advocates failed: no substantive responses.
instead of feeding raw scratchpad text to the judge.
Related MCP server: Peer Review MCP Server
Install
Requires Python 3.11+.
git clone https://github.com/rudysanchez-mfb/council-mcp.git
cd council-mcp
python -m venv .venv
source .venv/bin/activate
pip install -e .That installs the council-mcp console script. You can also run the module
directly:
python -m council_mcp.serverBoth speak MCP over stdio and print only JSON-RPC to stdout; diagnostics go to stderr.
Configuration
Set API keys in the environment of your MCP host (not in this repository). Only providers with a key present are usable; a missing key is reported as a per-provider error and the rest of the council continues.
Provider | Environment variable | Default model | Endpoint |
Moonshot / Kimi |
|
|
|
DeepSeek |
|
|
|
xAI |
|
|
|
Google Gemini |
|
| Google OpenAI-compatible endpoint |
Default lineup: Kimi, DeepSeek, Grok, Gemini (the first num_advocates are
used). DeepSeek is the default judge. See .env.example.
The provider registry lives in council_mcp/providers.py, so adding a model or
an OpenAI-compatible endpoint is a small local change.
MCP host configuration
Replace values as needed. Use an absolute path to the installed console script
if your host does not inherit your shell PATH (which council-mcp).
Claude Desktop
~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or
%APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"council": {
"command": "council-mcp",
"env": {
"KIMI_API_KEY": "…",
"DEEPSEEK_API_KEY": "…",
"XAI_API_KEY": "…",
"GEMINI_API_KEY": "…"
}
}
}
}Cursor
~/.cursor/mcp.json (or project .cursor/mcp.json):
{
"mcpServers": {
"council": {
"command": "council-mcp",
"env": {
"KIMI_API_KEY": "…",
"DEEPSEEK_API_KEY": "…",
"XAI_API_KEY": "…",
"GEMINI_API_KEY": "…"
}
}
}
}OpenCode
opencode.json:
{
"mcp": {
"council": {
"type": "local",
"command": ["council-mcp"],
"environment": {
"KIMI_API_KEY": "…",
"DEEPSEEK_API_KEY": "…",
"XAI_API_KEY": "…",
"GEMINI_API_KEY": "…"
},
"enabled": true
}
}
}Hermes Agent
~/.hermes/config.yaml:
mcp_servers:
council-mcp:
command: council-mcp
args: []
env:
KIMI_API_KEY: "…"
DEEPSEEK_API_KEY: "…"
XAI_API_KEY: "…"
GEMINI_API_KEY: "…"If your host does not support env in the server block, export the keys in the
environment that launches the host instead.
Tools
council_deliberate
Runs a deliberation and returns JSON.
Argument | Type | Default | Meaning |
| string | required | The question for the council. |
| integer (2–4) |
| How many advocate models to query. |
| boolean |
| Include each advocate's raw content in the result. |
Result fields include:
synthesis— the judge's answer.advocates— per-model content, latency, cost,degraded,degraded_reason.errors— providers that failed after retry/backfill.advocates_count,advocates_failed,degraded_count,contributors_count.retried,backfill_count,cost_total,latency_ms,synthesis_truncated.
Coverage, not capacity: advocates_count includes degraded responses. The
real breadth of a deliberation is
advocates_count - degraded_count (also exposed as contributors_count).
Callers should read the latter before claiming "N models agreed."
council_list_models
Lists the provider/model pairs currently registered.
Degraded response detection
classify_degraded() runs after reasoning wrappers are stripped and returns
(degraded, reason). A response is degraded when:
Empty body — nothing substantive after stripping wrappers.
Long reasoning preamble — ≥600 chars before the first required heading plus a chain-of-thought cue, or ≥2000 chars before it regardless.
Re-drafted answer — a required heading appears more than once alongside a chain-of-thought cue.
Missing structure — a ≥1500-char response missing at least two of the
VERDICT/REASONING/RISKSheadings, with at least one CoT cue.
Detection is deliberately structural first: the advocate prompt requires the three headings, so their absence is the primary signal. A small phrase list only corroborates; it never triggers a flag on its own. Short markerless answers and long but clean structured answers are never flagged.
Degraded responses are retried once and then replaced from the unused-provider pool when possible. Anything still degraded is excluded from the judge input and reported in the result.
Latency budget
The default budgets are tuned for hosts that cap a tool call at 300 seconds:
Constant | Value | Purpose |
| 110 s | Per-request read timeout. |
| 120 s | Hard wall-clock cap on the parallel advocate phase. |
| 90 s | Hard cap on the judge call. |
| 40 s | Total budget for sequential retry/backfill. |
| 4000 | Output cap; also a latency cap for reasoners. |
| 4096 | Judge output cap. |
Worst case is approximately HARD_CAP_S + RETRY_BUDGET_S + JUDGE_HARD_CAP_S
(≈250 s). If your host allows longer calls, tune the constants in
council_mcp/council.py.
Tests
The test suite makes no network calls and needs no API keys; provider calls are mocked.
python -m venv .venv
source .venv/bin/activate
pip install -e '.[dev]'
pytestCoverage includes strip_think_tags, the reasoning_content fallback, every
classify_degraded branch, retry/backfill accounting, the retry budget guard,
fail-closed behavior, and server argument forwarding.
License
Apache License 2.0. See LICENSE and NOTICE.
Copyright 2026 Rudy Sanchez.
Available Tools
2 toolscouncil_deliberateA
Run multi-model deliberation: several models answer a question in parallel, then a judge synthesizes the best answer. Use for high-stakes decisions, architecture questions, debugging strategy, or any question where hallucination risk matters.
Inputs: question (required), include_raw, num_advocates. Returns: result
council_list_modelsA
List the provider/model pairs available for deliberation.
Inputs: no named parameters. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 2 tools
The two tools serve clearly distinct purposes: one executes a deliberation, the other enumerates available models. There is no plausible way to confuse listing models with running a deliberation.
Both names use the same 'council_' prefix followed by a clear verb_noun pattern (council_deliberate, council_list_models). The convention is predictable and readable.
Two tools is thin for a server, though the domain (multi-model deliberation with a judge) is narrow enough that a discovery tool plus a single execution tool may genuinely be all that's needed. It sits at the borderline of minimal but defensible.
The surface covers the core operation and model discovery, but there is no way to inspect or retrieve prior deliberations, configure the judge/models per call, or check status of a long-running deliberation. These are notable gaps an agent could hit, though the primary workflow is usable.
Maintenance
Related MCP Connectors
Multi-LLM council: 25+ frontier models in parallel, consensus scoring, verdict-first code review.
AI routing, memory, guardrails, and governance. Routes across Claude, GPT, Gemini.
OpenAI-compatible BYOK LLM gateway: per-task routing, shadow evaluation, promotion gates.
Related MCP Servers
- AlicenseAqualityDmaintenanceEnables multi-LLM deliberation via a council of panel members, judges responses for consensus and contradictions, and returns analysis to the calling model.34MIT
- AlicenseNot gradedqualityBmaintenanceEnables querying multiple LLM providers in a weighted quorum to obtain a consensus answer with a certainty score. Supports cost tiers, visible degradation, and provider-agnostic configurations.Apache 2.0

tokonomix-council-mcpofficial
AlicenseAqualityCmaintenanceEnables multi-model consensus decision-making for high-stakes AI decisions, using independent expert models and a judge to surface disagreements and ground decisions.11121 npm1MIT- AlicenseNot gradedqualityDmaintenanceEnables Claude Code to send a prompt to multiple LLMs simultaneously and get a judge model's structured comparison of their responses. This surfaces consensus, contradictions, and blind spots, letting Claude write a more informed final answer.1MIT