Cohere Common Errors

Overview

Quick reference for real Cohere API v2 errors with exact messages, causes, and fixes.

Prerequisites

cohere-ai SDK installed
CO_API_KEY configured
Access to error logs

Error Reference

400 — Bad Request: Missing Required Field

CohereError: model is required

Cause: API v2 requires model for all endpoints (Chat, Embed, Rerank, Classify).

Fix:

// Wrong (v1 style)
await cohere.chat({ messages: [...] });

// Correct (v2)
await cohere.chat({ model: 'command-a-03-2025', messages: [...] });

400 — Embed: Missing embedding_types

CohereError: embedding_types is required for embed models v3 and higher

Fix:

await cohere.embed({
  model: 'embed-v4.0',
  texts: ['hello'],
  inputType: 'search_document',
  embeddingTypes: ['float'],  // Required for v3+
});

400 — Embed: Missing input_type

CohereError: input_type is required for embed models v3 and higher

Fix: Use one of: search_document, search_query, classification, clustering, image.

401 — Invalid API Token

CohereError: invalid api token

Cause: CO_API_KEY is missing, wrong, or revoked.

Fix:

# Verify key is set
echo $CO_API_KEY

# Test directly
curl -H "Authorization: Bearer $CO_API_KEY" \
  https://api.cohere.com/v2/chat \
  -H "Content-Type: application/json" \
  -d '{"model":"command-r7b-12-2024","messages":[{"role":"user","content":"hi"}]}'

429 — Rate Limit Exceeded

CohereError: You are using a Trial key, which is limited to N calls/minute

Rate limits by key type:

Key Type	Chat	Embed	Rerank	Other
Trial	20/min	5/min	5/min	1000/month
Production	1000/min	1000/min	1000/min	Unlimited

Fix:

import { CohereError } from 'cohere-ai';

try {
  await cohere.chat({ model: 'command-a-03-2025', messages: [...] });
} catch (err) {
  if (err instanceof CohereError && err.statusCode === 429) {
    // Back off and retry
    await new Promise(r => setTimeout(r, 60_000)); // wait 60s for trial keys
    // retry...
  }
}

400 — Classify: Too Few Examples

CohereError: each unique label requires at least 2 examples

Fix:

await cohere.classify({
  model: 'embed-english-v3.0',
  inputs: ['This product is amazing'],
  examples: [
    // Need at least 2 examples PER label
    { text: 'I love it', label: 'positive' },
    { text: 'Great product', label: 'positive' },
    { text: 'Terrible', label: 'negative' },
    { text: 'Worst ever', label: 'negative' },
  ],
});

400 — Rerank: Too Many Documents

CohereError: too many documents, max is 1000

Fix: Batch your documents:

async function batchRerank(query: string, docs: string[], topN = 10) {
  const BATCH = 1000;
  let allResults: any[] = [];

  for (let i = 0; i < docs.length; i += BATCH) {
    const batch = docs.slice(i, i + BATCH);
    const resp = await cohere.rerank({
      model: 'rerank-v3.5',
      query,
      documents: batch,
      topN,
    });
    allResults.push(
      ...resp.results.map(r => ({ ...r, index: r.index + i }))
    );
  }

  return allResults.sort((a, b) => b.relevanceScore - a.relevanceScore).slice(0, topN);
}

400 — Chat: response_format with documents

CohereError: response_format is not supported with documents or tools

Cause: response_format: { type: 'json_object' } cannot be combined with documents or tools.

Fix: Use either JSON mode OR document/tool mode, not both. For structured RAG output, parse the text response yourself.

500/503 — Server Error

CohereError: internal server error

Fix: Retry with exponential backoff. If persistent, check status.cohere.com.

CohereTimeoutError

CohereTimeoutError: Request timed out

Fix: The SDK has a default timeout. Increase it or reduce payload:

const cohere = new CohereClientV2({
  token: process.env.CO_API_KEY,
  timeoutInSeconds: 120, // default is lower
});

Diagnostic Commands

# Check Cohere service status
curl -s https://status.cohere.com/api/v2/status.json | jq '.status'

# Verify API key works
curl -s -o /dev/null -w "%{http_code}" \
  -H "Authorization: Bearer $CO_API_KEY" \
  https://api.cohere.com/v2/chat \
  -H "Content-Type: application/json" \
  -d '{"model":"command-r7b-12-2024","messages":[{"role":"user","content":"ping"}]}'

# Check installed SDK version
npm list cohere-ai 2>/dev/null || pip show cohere 2>/dev/null

Error Handling Wrapper

import { CohereError, CohereTimeoutError } from 'cohere-ai';

function diagnoseCohereError(err: unknown): string {
  if (err instanceof CohereTimeoutError) {
    return 'TIMEOUT: Increase timeoutInSeconds or reduce input size';
  }
  if (err instanceof CohereError) {
    switch (err.statusCode) {
      case 400: return `BAD_REQUEST: ${err.message}`;
      case 401: return 'AUTH: Check CO_API_KEY';
      case 429: return 'RATE_LIMIT: Back off or upgrade key';
      case 500: return 'SERVER: Retry later, check status.cohere.com';
      default:  return `UNKNOWN (${err.statusCode}): ${err.message}`;
    }
  }
  return `UNEXPECTED: ${String(err)}`;
}

Instructions

Classify the failure by status and operation, collect a redacted correlation ID and relevant limits/model metadata, then apply the smallest safe recovery path. Do not retry authentication, validation, policy, or contract errors blindly; preserve the last certified result and use bounded backoff only for explicitly retryable provider/network failures.

Error Handling

Failure class	Safe response
Authentication/authorization	Stop requests and repair or rotate the scoped secret through its owner.
Input/model/schema validation	Reject the request, correct the contract, and retest in staging.
Rate limit or timeout	Defer with bounded retry/backoff and protect queued work.
Provider outage or unknown fault	Preserve safe metadata, check status, and use the incident path.

Output

Return a classified retryable or terminal error with model/operation context, safe correlation data, recovery decision, and escalation owner. Exclude API keys, complete prompts, source documents, retrieved content, and personal data from logs, tickets, and generic error messages.

Examples

On a 429, retain the job ID and reschedule through the rate-limit policy; on a 401, halt the worker and verify the secret reference with its owner. For an invalid embed request, correct the input type in a fixture and validate it before allowing a production batch to run.

Resources

Next Steps

For comprehensive debugging, see cohere-debug-bundle.