generated: '2026-08-23'
method: searched
source: https://tech.kakaoent.com/ai/using-melon-mcp-server-en/
sources:
- https://tech.kakaoent.com/ai/using-melon-mcp-server-en/
- https://mcp.melon.com/mcp
limit_count: 0
note: >-
  No rate limits are published anywhere on the Kakao Entertainment or Melon surface. The Melon MCP
  integration guide has an "API Limitations" section, but it covers subscription gating and regional
  content restrictions rather than request quotas. No rate-limit response headers were observed on the
  live anonymous tools/list exchange, and no 429 semantics are documented. An honest zero.
limits: []
response_headers:
  observed: []
  probed_on: https://mcp.melon.com/mcp
  probe: POST tools/list (JSON-RPC 2.0), anonymous, HTTP 200
  headers_returned:
  - date
  - server
  - cache-control
  - content-type
  - x-accel-buffering
  - x-envoy-upstream-service-time
  - via
  - alt-svc
  note: >-
    No X-RateLimit-*, no RateLimit-*, no Retry-After. The response is served through istio-envoy, which
    can enforce limits without advertising them, so absence of headers is not proof of absence of
    limits - it is proof an agent gets no runtime signal.
exhaustion_status_code: undocumented
schema_level_bounds:
  note: >-
    The only quantitative ceilings the provider publishes are per-request page-size bounds declared in
    the tool input schemas, not rate limits. Recorded here because they are the only numbers an agent
    can plan against.
  maximums:
  - param: limit
    maximum: 100
  - param: size
    maximum_observed:
    - 30
    - 50
    - 100
  - param: start
    maximum_observed: 100
  - tool: recommend_similar_songs_by_dj_mallang
    ceiling: up to 30 similar songs returned per call, per the tool description