Skip to content

Enhanced ChatGPT Clone: Features Agents, MCP, Skills, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active

3.6 KB
Raw Blame
import { rateLimit } from 'express-rate-limit';
import {
  ViolationTypes,
  resolveMCPAppRateLimits,
  DEFAULT_MCP_APP_ADMISSION_REQUESTS_PER_MINUTE,
} from 'librechat-data-provider';
import type { AugmentedRequest, Store } from 'express-rate-limit';
import type { Request, RequestHandler, Response } from 'express';
import type { TCustomConfig } from 'librechat-data-provider';

type MCPAppRateLimitKind = 'resource' | 'toolCall';

type MCPAppRateLimitRequest = Request & {
  user?: { id?: string };
  config?: {
    config?: {
      rateLimits?: TCustomConfig['rateLimits'];
    };
  };
};

type ViolationDetails = {
  type: ViolationTypes;
  max: number;
  limiter: 'user';
  windowInMinutes: 1;
};

export interface MCPAppRateLimiterDependencies {
  store: Store;
  score?: number | string;
  logViolation: (
    request: Request,
    response: Response,
    type: ViolationTypes,
    details: ViolationDetails,
    score?: number | string,
  ) => Promise<void>;
}

export interface MCPAppAdmissionRateLimiterDependencies extends MCPAppRateLimiterDependencies {
  getLimit: () => number | undefined;
}

const limiters: Record<
  MCPAppRateLimitKind,
  {
    key: keyof ReturnType<typeof resolveMCPAppRateLimits>;
    message: string;
  }
> = {
  resource: {
    key: 'resourcesPerMinute',
    message: 'Too many app resource requests. Try again later',
  },
  toolCall: {
    key: 'toolCallsPerMinute',
    message: 'Too many app tool call requests. Try again later',
  },
};

export function createMCPAppRateLimiter(
  kind: MCPAppRateLimitKind,
  dependencies: MCPAppRateLimiterDependencies,
): RequestHandler {
  const definition = limiters[kind];
  return rateLimit({
    windowMs: 60_000,
    limit: (request) => {
      const config = (request as MCPAppRateLimitRequest).config?.config;
      return resolveMCPAppRateLimits(config?.rateLimits)[definition.key];
    },
    handler: async (request, response) => {
      const type = ViolationTypes.TOOL_CALL_LIMIT;
      const details: ViolationDetails = {
        type,
        max: (request as AugmentedRequest).rateLimit.limit,
        limiter: 'user',
        windowInMinutes: 1,
      };
      await dependencies.logViolation(request, response, type, details, dependencies.score);
      response.status(429).json({ message: definition.message });
    },
    keyGenerator: (request) => String((request as MCPAppRateLimitRequest).user?.id ?? ''),
    store: dependencies.store,
  });
}

/** Shared inexpensive admission bucket evaluated before request-scoped MCP policy/config work. */
export function createMCPAppAdmissionRateLimiter(
  dependencies: MCPAppAdmissionRateLimiterDependencies,
): RequestHandler {
  return rateLimit({
    windowMs: 60_000,
    limit: () => {
      try {
        const configured = dependencies.getLimit();
        return typeof configured === 'number' && Number.isSafeInteger(configured) && configured > 0
          ? configured
          : DEFAULT_MCP_APP_ADMISSION_REQUESTS_PER_MINUTE;
      } catch {
        return DEFAULT_MCP_APP_ADMISSION_REQUESTS_PER_MINUTE;
      }
    },
    handler: async (request, response) => {
      const type = ViolationTypes.TOOL_CALL_LIMIT;
      const details: ViolationDetails = {
        type,
        max: (request as AugmentedRequest).rateLimit.limit,
        limiter: 'user',
        windowInMinutes: 1,
      };
      await dependencies.logViolation(request, response, type, details, dependencies.score);
      response.status(429).json({ message: 'Too many app requests. Try again later' });
    },
    keyGenerator: (request) => String((request as MCPAppRateLimitRequest).user?.id ?? ''),
    store: dependencies.store,
  });
}