@pipeworx/mcp-openrouter 0.1.0

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Pipeworx
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,55 @@
+# mcp-openrouter
+
+OpenRouter Models MCP.
+
+Part of [Pipeworx](https://pipeworx.io) — an MCP gateway connecting AI agents to 873+ live data sources.
+
+## Tools
+
+| Tool | Description |
+|------|-------------|
+
+## Quick Start
+
+Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):
+
+```json
+{
+  "mcpServers": {
+    "openrouter": {
+      "url": "https://gateway.pipeworx.io/openrouter/mcp"
+    }
+  }
+}
+```
+
+Or connect to the full Pipeworx gateway for access to all 873+ data sources:
+
+```json
+{
+  "mcpServers": {
+    "pipeworx": {
+      "url": "https://gateway.pipeworx.io/mcp"
+    }
+  }
+}
+```
+
+## Using with ask_pipeworx
+
+Instead of calling tools directly, you can ask questions in plain English:
+
+```
+ask_pipeworx({ question: "your question about Openrouter data" })
+```
+
+The gateway picks the right tool and fills the arguments automatically.
+
+## More
+
+- [All tools and guides](https://github.com/pipeworx-io/examples)
+- [pipeworx.io](https://pipeworx.io)
+
+## License
+
+MIT

package/package.json

@@ -0,0 +1,20 @@
+{
+  "name": "@pipeworx/mcp-openrouter",
+  "version": "0.1.0",
+  "description": "OpenRouter Models MCP.",
+  "type": "module",
+  "main": "src/index.ts",
+  "types": "src/index.ts",
+  "keywords": ["mcp", "mcp-server", "model-context-protocol", "pipeworx", "openrouter"],
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/pipeworx-io/mcp-openrouter"
+  },
+  "scripts": {
+    "typecheck": "tsc --noEmit"
+  },
+  "devDependencies": {
+    "typescript": "^5.7.0"
+  }
+}

package/server.json

@@ -0,0 +1,18 @@
+{
+  "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
+  "name": "io.github.pipeworx-io/openrouter",
+  "title": "Openrouter",
+  "description": "OpenRouter Models MCP.",
+  "version": "0.1.0",
+  "websiteUrl": "https://pipeworx.io/packs/openrouter",
+  "repository": {
+    "url": "https://github.com/pipeworx-io/mcp-openrouter",
+    "source": "github"
+  },
+  "remotes": [
+    {
+      "type": "streamable-http",
+      "url": "https://gateway.pipeworx.io/openrouter/mcp"
+    }
+  ]
+}

package/src/index.ts

@@ -0,0 +1,311 @@
+interface McpToolDefinition {
+  name: string;
+  description: string;
+  inputSchema: {
+    type: 'object';
+    properties: Record<string, unknown>;
+    required?: string[];
+  };
+}
+
+interface McpToolExport {
+  tools: McpToolDefinition[];
+  callTool: (name: string, args: Record<string, unknown>) => Promise<unknown>;
+  meter?: { credits: number };
+  cost?: Record<string, unknown>;
+  provider?: string;
+}
+
+/**
+ * OpenRouter Models MCP.
+ *
+ * Keyless catalog of ~340 LLMs across every major provider (Anthropic, OpenAI,
+ * Google, Meta, Mistral, DeepSeek, xAI, and more) via the OpenRouter models API
+ * (openrouter.ai/api/v1). Live cross-provider PRICING, context windows, and
+ * capabilities — answers "which models exist, how much do they cost, how big is
+ * their context window". Pricing is normalized to $ per 1M tokens for
+ * readability. Keyless.
+ */
+
+
+const BASE = 'https://openrouter.ai/api/v1';
+const UA = 'pipeworx/1.0 (+https://pipeworx.io)';
+
+interface RawModel {
+  id: string;
+  canonical_slug?: string;
+  name?: string;
+  created?: number;
+  description?: string;
+  context_length?: number;
+  architecture?: {
+    modality?: string;
+    input_modalities?: string[];
+    output_modalities?: string[];
+    tokenizer?: string;
+  };
+  pricing?: {
+    prompt?: string;
+    completion?: string;
+    request?: string;
+    image?: string;
+    web_search?: string;
+    input_cache_read?: string;
+    input_cache_write?: string;
+  };
+  top_provider?: {
+    context_length?: number;
+    max_completion_tokens?: number;
+    is_moderated?: boolean;
+  };
+  supported_parameters?: string[];
+  knowledge_cutoff?: string | null;
+  hugging_face_id?: string | null;
+  per_request_limits?: unknown;
+}
+
+const tools: McpToolExport['tools'] = [
+  {
+    name: 'list_models',
+    description:
+      'List LLMs from the OpenRouter catalog (~340 models across all major providers) with normalized $/1M-token pricing, context windows, and modality. Filter by search, free-only, minimum context, or modality (e.g. "image" for vision models). Sorted cheapest-first by completion price. Keyless.',
+    inputSchema: {
+      type: 'object',
+      properties: {
+        search: {
+          type: 'string',
+          description: 'Case-insensitive substring match against model id, name, and description. e.g. "claude", "gpt-4o", "deepseek".',
+        },
+        free_only: {
+          type: 'boolean',
+          description: 'If true, only return models whose prompt price is $0 (free tier).',
+        },
+        min_context: {
+          type: 'number',
+          description: 'Minimum context_length in tokens, e.g. 100000 or 200000.',
+        },
+        modality: {
+          type: 'string',
+          description: 'Substring match on architecture.modality, e.g. "image" to find vision/multimodal models, "text->text" for text-only.',
+        },
+        limit: {
+          type: 'number',
+          description: 'Max models to return (default 25, max 60).',
+        },
+      },
+    },
+  },
+  {
+    name: 'get_model',
+    description:
+      'Get full details for one OpenRouter model by id — normalized $/1M pricing (prompt, completion, plus per-request/image/web-search costs), context window, modality, supported parameters, tokenizer, knowledge cutoff. e.g. id "anthropic/claude-sonnet-4.5". Keyless.',
+    inputSchema: {
+      type: 'object',
+      properties: {
+        id: {
+          type: 'string',
+          description: 'Exact model id, e.g. "anthropic/claude-sonnet-4.5", "openai/gpt-4o", "google/gemini-2.5-pro".',
+        },
+      },
+      required: ['id'],
+    },
+  },
+  {
+    name: 'compare_models',
+    description:
+      'Side-by-side comparison of pricing ($/1M tokens), context window, and modality for 2-5 OpenRouter models. Answers "is claude-sonnet-4.5 cheaper than gpt-4o". Keyless.',
+    inputSchema: {
+      type: 'object',
+      properties: {
+        ids: {
+          type: 'array',
+          items: { type: 'string' },
+          description: 'Array of 2-5 exact model ids, e.g. ["anthropic/claude-sonnet-4.5", "openai/gpt-4o"].',
+        },
+      },
+      required: ['ids'],
+    },
+  },
+];
+
+async function callTool(name: string, args: Record<string, unknown>): Promise<unknown> {
+  try {
+    switch (name) {
+      case 'list_models':
+        return listModels(args);
+      case 'get_model':
+        return getModel(args);
+      case 'compare_models':
+        return compareModels(args);
+      default:
+        return { error: `Unknown tool: ${name}` };
+    }
+  } catch (e) {
+    return { error: e instanceof Error ? e.message : String(e) };
+  }
+}
+
+/** Convert a per-token price string to $ per 1M tokens, rounded to 4dp. null if missing. */
+function perMillion(priceStr: string | undefined): number | null {
+  if (priceStr === undefined || priceStr === null || priceStr === '') return null;
+  const n = Number(priceStr);
+  if (!Number.isFinite(n)) return null;
+  return Math.round(n * 1_000_000 * 1e4) / 1e4;
+}
+
+function provider(id: string): string {
+  const i = id.indexOf('/');
+  return i > 0 ? id.slice(0, i) : id;
+}
+
+/** Map a raw model into a compact record (full=false) or a detailed one (full=true). */
+function mapModel(raw: RawModel, opts: { full?: boolean } = {}): Record<string, unknown> {
+  const pricing = raw.pricing ?? {};
+  const arch = raw.architecture ?? {};
+  const promptPer1m = perMillion(pricing.prompt);
+  const completionPer1m = perMillion(pricing.completion);
+  const isFree = Number(pricing.prompt ?? NaN) === 0;
+
+  if (!opts.full) {
+    return {
+      id: raw.id,
+      name: raw.name,
+      context_length: raw.context_length ?? null,
+      prompt_per_1m: promptPer1m,
+      completion_per_1m: completionPer1m,
+      is_free: isFree,
+      modality: arch.modality ?? null,
+      provider: provider(raw.id),
+    };
+  }
+
+  const tp = raw.top_provider ?? {};
+  let description = raw.description ?? '';
+  if (description.length > 500) description = `${description.slice(0, 500)}…`;
+
+  return {
+    id: raw.id,
+    name: raw.name,
+    description,
+    context_length: raw.context_length ?? null,
+    pricing: {
+      prompt_per_1m: promptPer1m,
+      completion_per_1m: completionPer1m,
+      input_cache_read_per_1m: perMillion(pricing.input_cache_read),
+      input_cache_write_per_1m: perMillion(pricing.input_cache_write),
+      request: pricing.request ?? null,
+      image: pricing.image ?? null,
+      web_search: pricing.web_search ?? null,
+    },
+    architecture: {
+      modality: arch.modality ?? null,
+      input_modalities: arch.input_modalities ?? null,
+      output_modalities: arch.output_modalities ?? null,
+      tokenizer: arch.tokenizer ?? null,
+    },
+    top_provider: {
+      context_length: tp.context_length ?? null,
+      max_completion_tokens: tp.max_completion_tokens ?? null,
+      is_moderated: tp.is_moderated ?? null,
+    },
+    supported_parameters: raw.supported_parameters ?? [],
+    knowledge_cutoff: raw.knowledge_cutoff ?? null,
+    hugging_face_id: raw.hugging_face_id ?? null,
+    created: raw.created ? new Date(raw.created * 1000).toISOString() : null,
+  };
+}
+
+let _cache: RawModel[] | null = null;
+/** Fetch + cache the models list within a single call. */
+async function orGet(): Promise<RawModel[]> {
+  if (_cache) return _cache;
+  const res = await fetch(`${BASE}/models`, {
+    headers: { Accept: 'application/json', 'User-Agent': UA },
+  });
+  if (!res.ok) throw new Error(`openrouter: ${res.status} ${(await res.text()).slice(0, 200)}`);
+  const body = (await res.json()) as { data?: RawModel[] };
+  _cache = Array.isArray(body.data) ? body.data : [];
+  return _cache;
+}
+
+async function listModels(args: Record<string, unknown>): Promise<unknown> {
+  const all = await orGet();
+
+  const search = typeof args.search === 'string' ? args.search.trim().toLowerCase() : '';
+  const freeOnly = args.free_only === true;
+  const minContext = typeof args.min_context === 'number' ? args.min_context : 0;
+  const modality = typeof args.modality === 'string' ? args.modality.trim().toLowerCase() : '';
+  let limit = typeof args.limit === 'number' ? Math.floor(args.limit) : 25;
+  if (!Number.isFinite(limit) || limit < 1) limit = 25;
+  if (limit > 60) limit = 60;
+
+  let filtered = all.filter((m) => {
+    if (search) {
+      const hay = `${m.id} ${m.name ?? ''} ${m.description ?? ''}`.toLowerCase();
+      if (!hay.includes(search)) return false;
+    }
+    if (freeOnly && Number(m.pricing?.prompt ?? NaN) !== 0) return false;
+    if (minContext && (m.context_length ?? 0) < minContext) return false;
+    if (modality && !(m.architecture?.modality ?? '').toLowerCase().includes(modality)) return false;
+    return true;
+  });
+
+  // Sort cheapest-first by completion price ($/1M); nulls last.
+  filtered = filtered.sort((a, b) => {
+    const av = perMillion(a.pricing?.completion);
+    const bv = perMillion(b.pricing?.completion);
+    if (av === null && bv === null) return 0;
+    if (av === null) return 1;
+    if (bv === null) return -1;
+    return av - bv;
+  });
+
+  const models = filtered.slice(0, limit).map((m) => mapModel(m));
+  return { total_available: all.length, count: models.length, models };
+}
+
+async function getModel(args: Record<string, unknown>): Promise<unknown> {
+  const id = typeof args.id === 'string' ? args.id.trim() : '';
+  if (!id) return { error: 'provide a model id, e.g. "anthropic/claude-sonnet-4.5"', id: args.id ?? null };
+
+  const all = await orGet();
+  const raw = all.find((m) => m.id === id);
+  if (!raw) return { error: 'model not found', id };
+  return mapModel(raw, { full: true });
+}
+
+async function compareModels(args: Record<string, unknown>): Promise<unknown> {
+  const ids = Array.isArray(args.ids) ? args.ids.filter((x): x is string => typeof x === 'string') : [];
+  if (ids.length < 2 || ids.length > 5) {
+    return { error: 'provide 2-5 model ids', ids };
+  }
+
+  const all = await orGet();
+  const byId = new Map(all.map((m) => [m.id, m]));
+
+  const comparison: Record<string, unknown>[] = [];
+  const notFound: string[] = [];
+  for (const id of ids) {
+    const raw = byId.get(id);
+    if (!raw) {
+      notFound.push(id);
+      continue;
+    }
+    const compact = mapModel(raw);
+    comparison.push({
+      id: compact.id,
+      name: compact.name,
+      context_length: compact.context_length,
+      prompt_per_1m: compact.prompt_per_1m,
+      completion_per_1m: compact.completion_per_1m,
+      modality: compact.modality,
+      is_free: compact.is_free,
+    });
+  }
+
+  const result: Record<string, unknown> = { count: comparison.length, comparison };
+  if (notFound.length) result.not_found = notFound;
+  return result;
+}
+
+export default { tools, callTool, meter: { credits: 1 } } satisfies McpToolExport;

package/tsconfig.json

@@ -0,0 +1,14 @@
+{
+  "compilerOptions": {
+    "target": "ES2022",
+    "module": "ESNext",
+    "moduleResolution": "bundler",
+    "strict": true,
+    "esModuleInterop": true,
+    "skipLibCheck": true,
+    "outDir": "dist",
+    "rootDir": "src",
+    "declaration": true
+  },
+  "include": ["src"]
+}