coherence-mcp-server
Enables AI agents to interact with the Coherence Network platform, allowing them to browse ideas, record contributions, and access governance features via natural language.
README
<!-- AUTO-GENERATED from README.template.md. Edit the template, not this file. -->
Coherence Network
An open intelligence platform that traces every idea from inception to payout — with fair attribution, coherence scoring, and federated trust.
Idea → Research → Spec → Implementation → Review → Usage → Payout
↑ ↓
└────────── coherence scores at every stage ─────────┘
Every stage is scored for coherence (0.0–1.0) — measuring test coverage, documentation quality, and implementation simplicity. Contributors are paid proportionally to the energy they invested and the coherence they achieved.
Come in
Here, anyone or anything can arrive: a person, an AI agent, a local model, a crawler, or a maintainer carrying questions from your own repo.
Point the session at this GitHub repository, npm package, API, or MCP server and ask:
What is alive here, and what can we contribute?
The agent can read the invitation, inspect ideas and specs, sense what is moving, and work on anything it feels ready to touch. It is invited to return what changed with sources, limits, and care.
This can also help your own repo. Ask your agent to receive the Coherence Network invitation, look at your local project, and adapt the practice there without pretending your repo is this organism.
Useful doors:
- Web: coherencycoin.com/come-in
- API:
GET https://api.coherencycoin.com/api/agent/invitation - CLI:
npx coherence-cli agent invite - MCP:
npx coherence-mcp-server, then callcoherence_agent_invitation
Quickstart (< 15 minutes)
git clone https://github.com/seeker71/Coherence-Network.git
cd Coherence-Network
make dev-setup # api venv + web npm install
make test # ~8s flow-centric suite
make api-dev # http://localhost:8000
make web-dev # http://localhost:3000
Your first 5 minutes
Option A — no install, just browse:
# See what ideas exist (live network, no key needed)
curl -s https://api.coherencycoin.com/api/ideas?limit=5 | python3 -m json.tool
# Check network health
curl -s https://api.coherencycoin.com/api/health | python3 -m json.tool
Option B — install the CLI:
npm i -g coherence-cli
coh status # network health, idea count, your identity
coh ideas # browse the portfolio ranked by ROI
coh idea <id> # deep-dive: scores, open questions, value gaps
Option C — give your AI agent access:
Add to your Claude/Cursor MCP config:
{
"mcpServers": {
"coherence-network": {
"command": "npx",
"args": ["coherence-mcp-server"]
}
}
}
Then ask your agent: "What ideas have the highest ROI right now?"
How to contribute
Every contribution — code, docs, review, design, community — is tracked and fairly attributed.
# Link your identity (37 providers: GitHub, Discord, Ethereum, Solana, ORCID, ...)
coh identity setup
coh identity link github your-handle
# Submit a new idea
coh share
# Record any contribution
coh contribute
# Or contribute via the API
curl -s https://api.coherencycoin.com/api/contributions/record \
-X POST -H "Content-Type: application/json" \
-d '{"provider":"github","provider_id":"your-handle","type":"code","amount_cc":5}'
Contribute to this repo
git clone https://github.com/seeker71/Coherence-Network.git
cd Coherence-Network
pip install -e api/.[dev]
python3 -m pytest api/tests/ -x -q # 813+ tests
The workflow is: Spec → Test → Implement → CI → Review → Merge. Specs live in specs/. Tests are written before implementation. Review is invited before merge when another set of eyes would improve care, clarity, or trust.
How to exchange value
Stake on ideas you believe in. Fork ideas to take them new directions. Trace the full value chain from spark to payout.
coh stake <idea-id> 10 # stake 10 CC on an idea
coh fork <idea-id> # fork and evolve it
# View the full value chain
curl -s https://api.coherencycoin.com/api/value-lineage/links?limit=5 | python3 -m json.tool
# Preview how payouts are distributed
curl -s https://api.coherencycoin.com/api/value-lineage/links/LINEAGE-ID/payout-preview \
-X POST -H "Content-Type: application/json" \
-d '{"total_value": 1000}' | python3 -m json.tool
How governance works
Open governance — anyone can propose changes, anyone can vote. Federated instances operate independently.
# See open governance proposals
curl -s https://api.coherencycoin.com/api/governance/change-requests | python3 -m json.tool
# See federated nodes and their capabilities
curl -s https://api.coherencycoin.com/api/federation/nodes | python3 -m json.tool
The five pillars
| Pillar | In practice |
|---|---|
| Traceability | Every unit of value is traceable from idea through spec, implementation, usage, and payout. Nothing is lost. |
| Trust | Coherence scores (0.0–1.0) replace subjective judgement with measurable quality. |
| Freedom | Fork any idea. Run your own node. Vote on governance. No gatekeepers. |
| Uniqueness | Every idea, spec, and contribution is uniquely identified, scored, and ranked. |
| Collaboration | Multi-contributor attribution with coherence-weighted payouts. Fair by design. |
Ecosystem
Every part of the network links to every other. Jump in wherever makes sense.
| Surface | What it is | Link |
|---|---|---|
| Web | Browse ideas, specs, contributors, and value chains visually | coherencycoin.com |
| API | Full OpenAPI surface — the engine behind everything | api.coherencycoin.com/docs |
| CLI | Terminal-first access — npx coherence-cli help or npm i -g coherence-cli && coh help |
npm: coherence-cli |
| MCP Server | Typed tool surface for AI agents (Claude, Cursor, Windsurf) | npm: coherence-mcp-server |
| OpenClaw Skill | Auto-triggers inside any OpenClaw instance | ClawHub: coherence-network |
| skills.sh | Portable agent skill directory (same SKILL.md as ClawHub) |
skills.sh — submit skills/coherence-network/ |
| askill.sh | Secondary skill index for discovery | askill.sh — submit skills/coherence-network/ |
| Join the Network | Run a node and contribute compute | JOIN-NETWORK.md |
Tech stack
- API: FastAPI (Python) —
api/ - Web: Next.js 15 + shadcn/ui —
web/— live at coherencycoin.com - Graph: Neo4j
- Relational: PostgreSQL (in-memory store in dev)
- CLI: Node.js —
cli/— npm: coherence-cli - MCP: Node.js —
mcp-server/— npm: coherence-mcp-server
Documentation
- Setup · Status · Deploy · Runbook
- Spec Coverage · Spec Tracking · Glossary
- Join the Network · Model Routing
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。