Allows using local Ollama models as sub-agents via an OpenAI-compatible endpoint, enabling offline or self-hosted model delegation for coding tasks.
Allows using OpenAI models as sub-agents through an OpenAI-compatible provider configuration, enabling delegated coding tasks to be executed by OpenAI models.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Let a strong AI master (Claude Code) hand coding work to cheaper sub-agents over MCP. Each sub-agent can only change the paths its role owns, so several can work in parallel without colliding.
Zero dependencies (Python 3.9+, stdlib only)
Any OpenAI-compatible model: DeepSeek (default), OpenAI, OpenRouter, Ollama
Roles, prompts and models live in one ensemble.json
Ownership is enforced, including through symlinks. Secrets and .git are never touched
┌──────────────────────────────┐
│ MASTER (Claude Code) │ plans, defines contracts, reviews, runs tests
└──────────────┬───────────────┘
│ MCP: delegate_task / delegate_parallel
┌──────────────▼───────────────┐
│ ensemble-mcp │ builds prompt, calls model, checks paths
└───────┬──────────┬───────────┘
│ │ │
┌───────▼───┐ ┌────▼──────┐ ┌─▼─────────┐
│ backend │ │ web │ │ tests │ sub-agents (cheap model)
│ src/ │ │ web/ │ │ tests/ │ each limited to its own paths
└───────────┘ └───────────┘ └───────────┘The master owns the shared interfaces (contracts) that no sub-agent may change, and reviews everything that comes back.
Related MCP server: agentmux
pip install git+https://github.com/VeeraPD/ensemble-mcp
cd your-project
ensemble-mcp init # creates ensemble.json
export DEEPSEEK_API_KEY=sk-... # or put it in ./.env (git-ignored)
ensemble-mcp check # verifies config, key and prompts
claude mcp add ensemble -- ensemble-mcp -c "$PWD/ensemble.json" serveCopy templates/CLAUDE.md into your project so the master knows the workflow, then ask Claude to build something.
For large projects, add a middle level: one area lead per part of the codebase. Each area lead is a Claude Code subagent with its own ensemble.json and MCP server.
┌────────────────────────────────────────────┐
│ LEVEL 1 MASTER │ cross-area contracts, final review
└───────────────┬────────────────┬───────────┘
│ │
┌─────────▼──────┐ ┌──────▼─────────┐
│ LEVEL 2 │ │ LEVEL 2 │
│ mobile lead │ │ web lead │ Claude Code subagents,
│ (subagent) │ │ (subagent) │ one per area
└───┬───────┬────┘ └───┬────────┬───┘
│ │ │ │
┌──▼──┐ ┌──▼───┐ ┌──▼──┐ ┌──▼───┐
│ ui │ │store │ │ api │ │front │ LEVEL 3 workers
└─────┘ └──────┘ └─────┘ └──────┘ (cheap models, owned paths)claude mcp add ensemble-mobile -- ensemble-mcp -c /abs/mobile/ensemble.json serve
claude mcp add ensemble-web -- ensemble-mcp -c /abs/web/ensemble.json serveRules:
Cross-area contracts live in a folder only the master edits.
Owned paths must not overlap between configs.
Each area lead reports a short summary upward; the master does not re-read every file.
Add this level only when one lead can no longer hold the whole codebase. Otherwise the single-level setup above is simpler and cheaper.
ensemble.json, paths relative to the file:
{
"provider": {
"base_url": "https://api.deepseek.com",
"api_key_env": "DEEPSEEK_API_KEY",
"models": ["deepseek-chat", "deepseek-reasoner"],
"default_model": "deepseek-chat"
},
"shared_context": ["docs/CONVENTIONS.md"],
"roles": {
"backend": { "owns": ["src/"], "instructions_file": "docs/agents/backend.md" },
"tests": { "owns": ["tests/"], "instructions": "Test through public APIs only." }
}
}Key | Meaning |
| Any OpenAI-compatible base URL. Ollama: |
| Allowed models. Use the default first, a stronger one only for steps that fail |
| Files added to every role's prompt (coding rules) |
| Files or folders the role may change |
| The role's prompt |
| Cap for parallel tasks (default 5) |
Optional: roles.<name>.context, max_context_bytes, provider.timeout, provider.max_retries, provider.max_tokens, audit_log.
Tool | Purpose |
| One role, one task: |
| Several tasks with distinct roles, run concurrently |
| Roles, owned paths, models |
A good task states the goal, the exact file paths, the interfaces to match and how to check the result, and passes the relevant contracts as context_files. Each call returns the files changed and any paths that were rejected for being outside the role.
ensemble-mcp init [dir] create ensemble.json
ensemble-mcp check validate setup
ensemble-mcp run ROLE "task" -f file run one task without MCP (--dry-run to preview)
ensemble-mcp serve MCP server on stdio (default)Model output is untrusted. A file is only saved if its path is inside the project, not a secret or .git, and under the role's owns (checked before and after following symlinks). Nothing the model returns is ever executed. Review every diff, and note that context files are sent to your model provider. See SECURITY.md and docs/ARCHITECTURE.md.
PYTHONPATH=src python3 -m unittest discover -s testsMIT. Not affiliated with Anthropic or DeepSeek.
This server cannot be deployed
Coordinate coding agents through MCP using existing AI plans, saved work, and independent checks.
Parallel coding agents claim a file before editing, so none overwrites another.
Coordination hub for AI coding agents: message teammates, ask humans, audit every event.
Human-agent interface for projects: you prioritize and decide, your AI agents execute. Over MCP.
