# MCP Server Search (`mcp.search`)

Searches public Model Context Protocol servers by capability or name and returns each with its description, repository, hosted endpoint or package, and version.

Docs: https://primaryapi.com/docs/api-reference/mcp/search

- Endpoint: `POST https://api.primaryapi.com/v1/run/mcp.search`
- Category: Knowledge & Data · Platform: `mcp` · Kind: search · Status: live
- Price: 1 credit per call. Every call costs the same, including calls that were retried automatically.
- Billing unit: per successful call
- Reliability: routed to a healthy upstream and retried automatically when an attempt fails; live 30-day status: `GET https://api.primaryapi.com/v1/status/apis?sku=mcp.search`
- Execution: synchronous (answers 202 only if it outlives the wait window)

## Input

| Field | Type | Required | Default | Constraints | Description |
|---|---|---|---|---|---|
| `query` | string | yes |  | max 100 chars | What the server should do or be called, e.g. "postgres" or "github issues" (3+ characters). |
| `remoteOnly` | boolean | no | false |  | Only servers with a hosted (remote) endpoint. |
| `includeDeprecated` | boolean | no | false |  | Also return servers their publisher marked deprecated. |
| `limit` | integer | no | 20 | 1–50 | Maximum number of results to return (1-50, default 20). |

Unknown fields are rejected with `400 invalid_input`. The routing controls `preferLatencyUnderMs` and `maxCredits` may be sent in the same JSON body.

## Output

When `output.found` is `true`, `output.data` has this shape:

| Field | Type | Description |
|---|---|---|
| `servers` | object[] | Best matches first, then by name. |
| `servers[].name` | string | Reverse-DNS server name, e.g. "io.github.owner/server". |
| `servers[].title` | string \| null |  |
| `servers[].description` | string \| null |  |
| `servers[].version` | string \| null |  |
| `servers[].repository` | string \| null | Source repository URL. |
| `servers[].websiteUrl` | string \| null |  |
| `servers[].remotes` | object[] | Hosted endpoints you can connect to. |
| `servers[].remotes[].type` | string | streamable-http or sse. |
| `servers[].remotes[].url` | string |  |
| `servers[].packages` | object[] | Installable packages (run locally). |
| `servers[].packages[].registryType` | string | npm, pypi, oci, nuget, mcpb, … |
| `servers[].packages[].identifier` | string |  |
| `servers[].packages[].version` | string \| null |  |
| `servers[].packages[].transport` | string \| null |  |
| `servers[].status` | string | active or deprecated. |
| `servers[].publishedAt` | string \| null |  |
| `servers[].updatedAt` | string \| null |  |
| `servers[].score` | integer | Query words found in name, title or description (higher ranks first). |
| `total` | integer | Matching servers before `limit`. |
| `nextCursor` | string \| null |  |

When the thing you asked about does not exist (or a search matches nothing) the call still succeeds: `output.found` is `false`, `output.data` is `null` and `output.reason` says why (`not_found`, `private`, `no_results`, …). Misses are free unless the notes below say misses are billed.

## Example request

```bash
curl -sS -X POST "https://api.primaryapi.com/v1/run/mcp.search" \
  -H "Authorization: Bearer $PRIMARYAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query":"postgres"}'
```

## Example response

_Values are placeholders that follow the output schema; `credits` is what this call would cost at the usual price._

```json
{
  "output": {
    "found": true,
    "data": {
      "servers": [
        {
          "name": "string",
          "title": "string",
          "description": "string",
          "version": "string",
          "repository": "string",
          "websiteUrl": "https://example.com",
          "remotes": [
            {
              "type": "string",
              "url": "https://example.com"
            }
          ],
          "packages": [
            {
              "registryType": "string",
              "identifier": "string",
              "version": "string",
              "transport": "string"
            }
          ],
          "status": "string",
          "publishedAt": "string",
          "updatedAt": "string",
          "score": 0
        }
      ],
      "total": 0,
      "nextCursor": null
    }
  },
  "provider": "PrimaryAPI",
  "credits": 1,
  "items": 1,
  "requestId": "req_01JBXK6Y2N9V8T4R3Q2P1M0A7Z",
  "nextCursor": null
}
```

## Pricing and reliability

- Usual price: 1 credit per call
- Price ceiling: the usual price, even when a call is retried. A quote’s `maxCredits` is the most a given call can cost; the `maxCredits` routing control caps it per call
- Live 30-day status, success rate, uptime and latency: `GET https://api.primaryapi.com/v1/status/apis?sku=mcp.search`

## Errors to handle

`invalid_input`, `unauthorized`, `insufficient_balance`, `key_limit_reached`, `trial_cap_reached`, `forbidden`, `not_found`, `idempotency_conflict`, `rate_limited`, `all_sources_failed`, `service_unavailable`, `timeout`. See https://primaryapi.com/docs/errors.

## Notes

Candidates are servers whose name contains your query or one of its words ("postgres" finds "io.github.x/postgres-mcp"); they are ranked by how many of your words appear in the name, title and description. Only the latest version of each server is returned. No match is a free miss (reason no_results). Listings are metadata published by server authors: evaluate a server before you connect it.

