cohere-rate-limits

v2026.09.24

Implement Cohere endpoint-aware throttling, bounded retry, and queue backpressure from current limits and observed responses. Use when handling 429s or planning throughput. Trigger with "Cohere rate limit", "Cohere throttling", or "Cohere 429".

GitHub
Install command
npx skhub add jeremylongshore/cohere-rate-limits
Markdown
SKILL.md

Cohere Rate-Limit Control

Overview

Treat limits as a live per-key, endpoint, and model contract rather than a single permanent requests-per-minute number.

Prerequisites

  • The target repository, runtime, environment, and accountable owner
  • An approved Cohere team and key for any live verification
  • Current quality, security, privacy, capacity, and change-control requirements

Tool Discipline

Use Read, Glob, and Grep to inspect code, configuration, and evidence. Use WebFetch only for current Cohere primary documentation. Use Write or Edit only when the user requested implementation and the exact target files are known; never write credentials or customer content.

Current Contract

  • The reviewed page lists Chat limits per model; several newer variants require contacting sales for production capacity.
  • On the review date, Embed was expressed in inputs per minute, while Rerank was expressed in requests per minute.
  • Trial usage includes a monthly cap, and production arrangements can vary.
  • A successful load plan must measure actual account behavior and preserve headroom instead of copying a static table.

Authentication

Use an environment-specific key injected from an approved secret manager. Never print, persist, commit, or place CO_API_KEY in an example. Confirm access with the least costly bounded operation appropriate to the task, and treat key creation, rotation, revocation, role changes, and production-capacity requests as owner-approved actions.

Instructions

  1. Identify key type, endpoint, resolved model, request batch size, and documented current limit.
  2. Convert limits into request, input, and token budgets as applicable.
  3. Apply a shared queue with bounded concurrency, jittered backoff, and a total retry deadline.
  4. Honor provider retry guidance when present and stop retrying deterministic client errors.
  5. Load test below the approved ceiling with non-sensitive fixtures and record saturation behavior.
  6. Alert on sustained throttling, queue age, dropped work, and budget exhaustion.

Approval Boundaries

Do not expose or rotate keys, change Cohere Team roles, accept commercial terms, enable sensitive production data, increase spend or capacity, switch production models, send a support bundle, or execute model-proposed side effects without the accountable owner's approval. Keep diagnosis read-only unless implementation was requested.

Output

Return the resolved API and model contract, files or settings inspected, evidence collected, validation result, remaining risk, owner, and rollback or next action. Redact keys, authorization headers, prompts, retrieved documents, embeddings, customer identifiers, and unrestricted environment output.

Error Handling

ConditionResponse
Repeated 429Reduce concurrency and inspect key/model-specific capacity.
Queue growthApply admission control instead of hiding overload with retries.
Mixed endpointsUse separate budgets because units and limits differ.
Unknown limitRun a conservative probe or contact Cohere; do not invent capacity.

Examples

Use this compact handoff shape to keep the selected scope, validation evidence, and operational result reviewable.

Input:

endpoint=rerank; key=production; model=resolved; workload=interactive

Expected handoff:

concurrency=measured; retry=bounded-jitter; headroom=20%; alerts=configured

Resources

Discovery
Tags

No tags published for this skill.

Version
Latest version metadata

Version

v2026.09.24

Published

Sep 24, 2026

Category

Uncategorized

License

MIT

Source path

skills/.curated/cohere-rate-limits

Default branch

main

Latest commit

e5a6c3b

Tree SHA

c2dc8e8