dsh-throttle

dsh-throttle

English | 中文

DeepSeek Harness rate-limiting plugin — a transparent API proxy for the Harness ecosystem, reusing 429-throttle-mcp core logic.

---

What is it

Many free LLM APIs (Grok, Gemini, Dots, etc.) only allow ~30 calls per minute. During long-running tasks, agents easily exceed the quota and get rejected with HTTP 429.

dsh-throttle provides a transparent rate-limiting layer purpose-built for the DeepSeek Harness ecosystem:

Agent → call_api tool → rate limiter → actual API request → result + usage snapshot

It shares the same rate-limiter.js core as 429-throttle-mcp, differing only in the communication protocol — MCP stdio JSON-RPC vs. DSH Plugin API.

---

Core Parameters

Config source: DSH config or dsh-manifest.json.

---

Exposed Tools

call_api

Send an HTTP request through the rate-limiting proxy.

If rate-limited, returns a RATE_LIMIT_EXCEEDED error with retryAfterSeconds suggesting how long to wait.

get_rate_limit_status

Query current rate-limit usage. Returns used/remaining calls and tokens.

set_rate_limit

Dynamically adjust rate-limit parameters (takes effect immediately, no restart needed).

---

Installation

npm install dsh-throttle

Add to your DSH config:

{
  "plugins": {
    "dsh-throttle": {
      "maxCalls": 30,
      "maxTokens": 750000
    }
  }
}

Load the plugin:

dsh plugin load dsh-throttle

---

Relationship with 429-throttle-mcp

---

SEO Keywords

429 error, anti 429, DSH rate limit, DeepSeek Harness plugin, LLM per-minute call limit, free LLM rate limiting, agent batch calls triggering 429, RPM, TPM, rate limiter, quota guard, MCP proxy, throttle, LLM API quota, cop, HTTP 429, Too Many Requests, rate limiting, token bucket, sliding window

---

License

MIT