Beam logoBeam

Rate Limits And Reliability

Beam respects Figma limits and permissions.

Caching, snapshots, request deduplication, retry metadata, and cached fallback behavior reduce repeated Figma calls without claiming unlimited upstream access.

When Figma returns a rate limit, Beam should report the endpoint, retry timing when available, whether cached data exists, and whether the agent can proceed from a snapshot.

If a cached snapshot is good enough, Beam can continue with a build-readiness review. If the snapshot is incomplete, Beam asks for specific user evidence instead of guessing silently.

Draft files, private files, missing node access, and limited token scopes are product states Beam must explain clearly rather than masking them as generic request failures.

Run Sequence

  1. Make one deliberate Figma fetch instead of repeated blind retries.
  2. Cache raw node data and rendered images from successful fetches.
  3. On a 429, read retry metadata and report when Beam can retry.
  4. Offer cached snapshot fallback when available.
  5. If cached evidence is insufficient, ask for the smallest useful user evidence.

Expected Output

A rate-limit response should name the endpoint, retry timing when available, cache fallback status, snapshot age, and whether the build is ready, degraded, or blocked.

Files And Artifacts

  1. .beam/cache/raw/
  2. .beam/cache/images/
  3. .beam/cache/snapshots/
  4. .beam/cache/logs/

Next Step

Proceed from cache when confidence is high enough, or wait for Figma retry timing when fresh data is required.

Recovery Paths

  1. If the file is private, fix sharing rather than retrying.
  2. If the token lacks access, log in with an account that can open the file.
  3. If repeated fetches cause limits, reuse snapshots and avoid unnecessary refreshes.