# OriginGrid

> AI Agent Source Registry. 288K+ curated sources for agentic search and discovery.

- name: net.origingrid/origingrid-mcp
- version: 1.0.1
- connection class: unknown (unknown)
- trust: 60/100 flags: unreachable
- quality: 88/100 (good)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/net.origingrid%2Forigingrid-mcp
- tools: https://api.protogrid.dev/v1/servers/net.origingrid%2Forigingrid-mcp/tools

**Not classified yet.** The remote endpoints have not been probed successfully, or the listing has nothing to connect to.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "origingrid-mcp": {
      "type": "http",
      "url": "https://www.origingrid.net/mcp"
    }
  }
}
```


## Trust

- hygiene: 100
- liveness: 0
- freshness: 85
- provenance: 75
- drivers: +repository +dns-namespace -unreachable(6 failures) ~updated-83d-ago
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 88/100 (good)

### protocol: 100 (weight 25)

- pass `protocol.modern`: supports 2026-07-28 (server/discover)
- pass `protocol.stateless`: answers without a session
- pass `protocol.transport`: streamable HTTP
- pass `protocol.list_ttl`: tool list is cacheable (ttlMs 300000)

### authorization: n/a (weight 15)

- n/a `auth.prm`, `auth.as_metadata`, `auth.cimd`: no remote uses OAuth
- n/a `auth.secret_in_url`: no templated URL

### tool hygiene: 71 (weight 30)

- pass `tools.descriptions`: every tool has a description
- pass `tools.description_length`: descriptions are concise
- pass `tools.schemas`: every tool has an object input schema
- fail `tools.annotations`: no tool declares readOnlyHint or destructiveHint, so clients cannot tell safe calls from risky ones
- fail `tools.directory_hints`: no tool declares all four of readOnlyHint, destructiveHint, idempotentHint and openWorldHint; missing: origingrid_pending_contributions (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); origingrid_discover (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); origingrid_plan (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); …
- pass `tools.token_cost`: about 2,857 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 100 (weight 15)

- pass `stability.changes`: no tool changes in 30 days
- pass `stability.rug_pull`: no tool changed its meaning under the same name

### dependencies: n/a (weight 15)

- n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package

- tools: 19, about 2,857 tokens to load them all
- tool set hash: f24e11b69bc68d4c84059cb15627c52a
- badge: [![protogrid quality](https://protogrid.dev/badge/net.origingrid/origingrid-mcp.svg)](https://protogrid.dev/servers/net.origingrid/origingrid-mcp)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

_No tool definition changes recorded._

## Remotes

- https://www.origingrid.net/mcp (streamable-http, auth none, reachable false, uptime30d 0.925)

## Packages

_none_

## Tools (19)

- `origingrid_approve_contribution`: Approve a pending source-card contribution (admin only; requires admin_token when ORIGINGRID_ADMIN_TOKEN is set)
- `origingrid_contribute`: Submit a new URL as a source-card candidate (anonymous, pending review). [ASRP: Use AFTER finding a URL that is useful but not yet in the registry.]
- `origingrid_discover`: Discover first-hand information sources for a topic from the OriginGrid local registry — no external search engine involved. Optionally narrow by topic and/or region. Follow up with origingrid_fetch(url) for live content or origingrid_get_source for details. [ASRP: Use FIRST for broad topic discovery. Follow with origingrid_plan for multi-step fetch or origingrid_fetch for single URL.]
- `origingrid_execute_batch`: Execute multiple search plan steps in batch. Accepts `steps` (array of step objects from origingrid_plan) and optional `parallel_groups` (array of arrays of step indices). Steps within the same parallel_group run concurrently on the server — the agent doesn't need to manage concurrency. Supports per-step timeout, automatic retry (default 1), and error isolation (one step failing doesn't affect others). Concurrency is limited to 8 by default for VPS memory safety. Returns structured results with elapsed timing per step. Combine with origingrid_plan: plan → execute_batch → agent reads the results. [ASRP: Call AFTER origingrid_plan. Steps in same parallel_group run concurrently.]
- `origingrid_feedback`: Submit feedback (rating 1-5) for a source returned in a previous search, using that search's `search_id` (from origingrid_search response). Feeds source quality signals; rating 3 = neutral, no positive/negative bias.
- `origingrid_fetch`: Fetch first-hand content from a URL. Returns `structuredContent.data` with {title, content, url, cache_hit, fetched_at}. Read `.data.content` directly — the metadata and content are separated. Pass `delivery_level=excerpt` (default) for ~300 chars, `full` for complete content. [ASRP: Call AFTER origingrid_search. Loop over sources[] and fetch each source.url.]
- `origingrid_get_source`: Get details for a single information source by its source_id
- `origingrid_list`: List available information sources, optionally filtered by topic and/or region. Supports `limit`/`offset` pagination; the response reports the **total** number of matches so you can page through all results. [ASRP: Browse available sources by category. Use for exploration before planning.]
- `origingrid_pending_contributions`: List pending source-card contributions awaiting admin review (admin only; requires admin_token when ORIGINGRID_ADMIN_TOKEN is set). Review loop: list pending → origingrid_approve_contribution(id)
- `origingrid_plan`: Generate an executable search plan for an AI agent. Returns a structured JSON plan with ordered steps, each containing an executable URL with method, expected latency, difficulty, and priority. The agent executes the steps itself — no search-engine caching. Steps within the same parallel_group can be fetched concurrently. Requires browser for yellow-difficulty sources. [ASRP: Call AFTER search/discover. Pass the returned steps to origingrid_execute_batch for concurrent execution.]
- `origingrid_reject_contribution`: Reject a pending source-card contribution (admin only; requires admin_token when ORIGINGRID_ADMIN_TOKEN is set). Optionally pass `reason`. Review loop: list pending → origingrid_approve_contribution(id) / origingrid_reject_contribution(id)
- `origingrid_resolve`: Resolve a natural-language query into directly executable first-hand source URLs. Returns structured JSON with executable_url that the agent opens locally. Supports searchable sources (API/search URLs with query parameters) and navigable sources (key pages like about/products/docs/pricing). Does NOT proxy-fetch or submit URLs to general-purpose search engines. [ASRP: Lightweight alternative to plan+execute_batch. Use for direct URL resolution without multi-step plans. For complex queries, use origingrid_plan instead.]
- `origingrid_search`: Search for AI agent information sources. Returns `structuredContent.sources[]` with {source_id, name, url, region, difficulty, score}. Read the array directly — no Markdown parsing. Then call origingrid_fetch for each source's URL to get actual content. [ASRP: Use FIRST to find sources. Then fetch each source.url.]
- `origingrid_search_by_region`: Search sources filtered by region (cn/global/academic/etc) and optionally by query. [ASRP: Use for region-specific discovery. Narrower than origingrid_search, use when you know the target region.]
- `origingrid_search_session_create`: Create a search session that persists across multiple tool calls. Generates a plan and returns a session_id. Call origingrid_search_session_execute with this session_id to execute steps group by group. [ASRP: Use INSTEAD of origingrid_plan+execute_batch when you need persistent session state across calls.]
- `origingrid_search_session_execute`: Execute the next unexecuted step group in a search session. Pass the session_id from origingrid_search_session_create. Returns results for the current group and reports remaining groups. [ASRP: Call AFTER origingrid_search_session_create. Repeat until remaining_groups is 0.]
- `origingrid_search_session_status`: Get the current status of a search session. Returns query, depth, completed/total groups, progress percentage, and timestamps. [ASRP: Use to check session progress between execute calls.]
- `origingrid_search_similar`: Find sources similar to a known source. [ASRP: Use AFTER finding a relevant source to expand your search scope before calling origingrid_plan.]
- `origingrid_stats`: Get OriginGrid registry statistics