api
Anthropic API
17 verified traps, most recently checked first.
- temperature or top_p returns 400, or a TypeError in the Python SDK, on Claude Opus 4.7 and later
400 invalid_request_error naming temperature, top_p, or top_k
Anthropic API · verified 2026-10-15 · 0 confirmations
- tool_choice type "tool" or "any" returns 400 on Claude Opus 5.5, Sonnet 5.5, and Fable 5.1
tool_choice: type "tool" and "any" are not supported for this model.
Anthropic API · verified 2026-10-15 · 0 confirmations
- 400 "thinking.type.enabled" is not supported after switching to a newer Claude model
400 invalid_request_error: "thinking.type.enabled" is not supported for this model
Anthropic API · verified 2026-10-15 · 0 confirmations
- 400 Invalid signature in thinking block: 'The block is bound to a different conversation'
messages.N.content.M: Invalid `signature` in `thinking` block. The block is bound to a different conversation.
Anthropic API · verified 2026-10-15 · 0 confirmations
- 400 "thinking.type.disabled" is not supported when turning thinking off on Opus 5.5 or Sonnet 5.5
"thinking.type.disabled" is not supported for this model. Use "thinking.type.adaptive" and "output_config.effort" to control thinking behavior.
Anthropic API · verified 2026-10-15 · 0 confirmations
- 429 rate_limit_error that retries never clear because the monthly API spend limit was reached
429 rate_limit_error: You have reached your API usage limits
Anthropic API · verified 2026-10-15 · 0 confirmations
- Assistant prefill returns 400 on Claude Opus 4.6 and later models
400 invalid_request_error when the messages array ends with an assistant turn
Anthropic API · verified 2026-10-15 · 0 confirmations
- Calling the Claude API from the browser fails with a CORS error or a browser-like environment error
It looks like you're running in a browser-like environment. This is disabled by default, as it risks exposing your secret API credentials to attackers.
Anthropic API · verified 2026-10-15 · 0 confirmations
- Prompt cache misses after changing tool_choice, images, effort, or thinking settings between calls
cache_read_input_tokens drops to 0 on a turn where only tool_choice changed
Anthropic API · verified 2026-10-15 · 0 confirmations
- Prompt caching silently does nothing: cache_creation_input_tokens and cache_read_input_tokens stay 0
usage.cache_creation_input_tokens: 0 and usage.cache_read_input_tokens: 0 on every call
Anthropic API · verified 2026-10-15 · 0 confirmations
- SDK throws 'Streaming is required for operations that may take longer than 10 minutes' before sending
Streaming is required for operations that may take longer than 10 minutes. See https://github.com/anthropics/anthropic-sdk-typescript#long-requests for more details
Anthropic API · verified 2026-10-15 · 0 confirmations
- Structured outputs return 400 'Schema is too complex for compilation' or 'compiled grammar is too large'
The compiled grammar is too large, which would cause performance issues. Simplify your tool schemas or reduce the number of strict tools.
Anthropic API · verified 2026-10-15 · 0 confirmations
- Message Batches results come back in a different order than the requests were submitted
Answers are attached to the wrong input rows
Anthropic API · verified 2026-10-15 · 0 confirmations
- Claude with web search narrates before its JSON, so a strict JSON parser reports nothing
JSON.parse fails or returns an empty list
Anthropic API · verified 2026-09-24 · 0 confirmations
- Claude response is empty with stop_reason max_tokens and no error
stop_reason: max_tokens with zero text blocks
Anthropic API · verified 2026-08-02 · 0 confirmations
- content[0].text is undefined when the model returns a thinking block first
Cannot read properties of undefined
Anthropic API · verified 2026-07-18 · 0 confirmations
- Long Claude request fails with 'Premature close' on an old SDK
Premature close
Anthropic API · verified 2026-07-13 · 0 confirmations
Give your coding agent these traps before it hits them
claude mcp add --transport http footgun https://getfootgun.com/api/mcp