mcpField
O
ophir

ophir

Local
github.com/Ophir-ProtocolAI & Agents0Updated 34w ago
Repository

About ophir

ophir is an MCP server for AI agents to find AI service providers, negotiate terms, and manage agreements through the Ophir protocol. It also supports SLA monitoring and disputes.

Key features

  1. 01Discover providers, prices, and reputation
  2. 02Request and rank quotes from providers
  3. 03Accept quotes and create signed agreements
  4. 04Review agreements and SLA status
  5. 05File SLA disputes with cryptographic evidence

Use cases

  1. 01Find inference providers above a reputation threshold
  2. 02Negotiate service pricing within a per-request budget
  3. 03Check agreements against SLA targets
  4. 04Dispute latency violations

MCP support

The MCP server exposes Ophir tools for provider discovery, quote negotiation and acceptance, agreement checks, and SLA disputes. It connects to the public Ophir registry.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "ophir-ai-creator-ophir": {
      "command": "npx",
      "args": [
        "-y",
        "@ophirai/mcp-server"
      ]
    }
  }
}
Run commandnpx -y @ophirai/mcp-server

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npm@ophirai/mcp-server0.2.1

Tools

7

Saved tool list · updated 3w ago

ParameterTypeRequiredDescription
categorystringNoService category to filter by (e.g. "inference", "translation", "code_review"). Omit to list all providers.
min_reputationnumberNoMinimum reputation score (0-100) to filter sellers. Only providers at or above this score are returned.

Details

From the project README

@ophirai/mcp-server

MCP server that gives your AI agent the ability to discover, negotiate with, and manage AI service providers through the Ophir protocol.

No API keys required — connects to the public Ophir registry.

Installation

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "ophir": {
      "command": "npx",
      "args": ["@ophirai/mcp-server"]
    }
  }
}
Cursor

Add to .cursor/mcp.json in your project root:

{
  "ophir": {
    "command": "npx",
    "args": ["@ophirai/mcp-server"]
  }
}
Windsurf

Add to your Windsurf MCP config:

{
  "mcpServers": {
    "ophir": {
      "command": "npx",
      "args": ["@ophirai/mcp-server"]
    }
  }
}
Cline

Add to your Cline MCP settings:

{
  "mcpServers": {
    "ophir": {
      "command": "npx",
      "args": ["@ophirai/mcp-server"]
    }
  }
}
Any MCP client (npx)
npx @ophirai/mcp-server
Smithery
smithery install ophir-negotiate

Configuration

Environment VariableDescriptionDefault
OPHIR_REGISTRY_URLOphir registry endpointhttps://registry.ophir.ai/v1
OPHIR_SELLERSComma-separated seller endpointsAuto-discovered
OPHIR_BUYER_ENDPOINTLocal buyer agent endpointhttp://localhost:3001

What can it do?

ophir_discover — Find providers

"Find me inference providers with a reputation score above 80."

Returns a list of providers with their services, pricing, and reputation scores:

{
  "providers": [
    {
      "agent_id": "provider-abc",
      "endpoint": "https://provider-abc.ophir.ai",
      "services": [{ "category": "inference", "base_price": "0.002", "currency": "USDC", "unit": "request" }],
      "reputation": { "score": 92, "total_agreements": 1847 }
    }
  ],
  "total": 3
}
ophir_negotiate — Full negotiation flow

"Negotiate inference service with a max budget of 0.005 USDC per request."

Sends an RFQ to all matching providers, collects quotes, ranks them, and returns the best option:

{
  "best_quote": {
    "seller": "provider-abc",
    "price": "0.0018",
    "currency": "USDC",
    "unit": "request",
    "sla": [
      { "metric": "latency_p99", "target": "500ms" },
      { "metric": "uptime", "target": "99.9%" }
    ]
  },
  "total_quotes": 5,
  "rfq_id": "rfq_01HXYZ..."
}
ophir_accept_quote — Accept a quote

"Accept quote qt_abc123 from RFQ rfq_01HXYZ."

Creates a signed agreement with the selected provider, locking in the negotiated terms.

ophir_check_agreement — Monitor SLA compliance

"Check if agreement agr_xyz is meeting its SLA targets."

Returns current SLA metric observations and compliance status:

{
  "agreement_id": "agr_xyz",
  "sla_metrics": [
    { "name": "latency_p99", "target": "500ms", "comparison": "lte" },
    { "name": "uptime", "target": "99.9%", "comparison": "gte" }
  ],
  "compliance_status": "monitoring_available"
}
ophir_list_agreements — View active agreements

"Show me all my active provider agreements."

Lists all agreements with their pricing, status, and SLA terms.

ophir_dispute — File an SLA dispute

"File a dispute against agreement agr_xyz for latency violations."

Submits a dispute with cryptographic evidence of SLA violations, triggering the protocol's dispute resolution process.

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.