Anthropic
Call Claude models from an automation - generate text, extract schema-shaped data, and measure prompts - on your own API key.
Connect an Anthropic workspace with an API key to use Claude models as automation tools: generate text from a prompt, extract data that matches a JSON Schema, and count a prompt's input tokens for a specific model before sending it. List models is the source of truth for which model ids this key can reach, along with each one's context window and output ceiling. This is separate from the model that runs your automation - that is configured under Settings → AI settings → AI providers, and per agent under Setup → AI → Model & quality, on the organization's own keys. Sampling parameters (temperature, top_p) are not offered because current Claude models reject them; reasoning depth is set with effort instead.
Connect
| Credential field | Required | Where it comes from |
|---|---|---|
API key (API_KEY) | Yes | From platform.claude.com → API keys (starts with sk-ant-). It carries the model access, rate limits and billing of the workspace that owns it. |
| Setting | Required | What it is |
|---|---|---|
Default model (DEFAULT_MODEL) | No | Model id used when a step names none, e.g. claude-sonnet-5. Leave blank to use this integration's built-in default. A step can always override it. |
Tools
| Tool | Access | What it does |
|---|---|---|
List models (yekar.anthropic.list-models) | Read | The Claude models this API key can reach, newest first, with each one's context window and output ceiling. This is the source of truth for the model argument on the other tools: model access is granted per workspace, so a model in Anthropic's documentation is not necessarily enabled for this key. Note that ids carry no date suffix - claude-sonnet-5, not claude-sonnet-5-20260101. |
Generate text (yekar.anthropic.generate-text) | Read | Send a prompt to a Claude model and return its answer. One request, one answer - this tool holds no conversation state, so anything the model needs to know belongs in prompt or system. Use it for a second opinion from a different model family, for work a specific Claude model is better at, or wherever an automation needs Claude text specifically. For a result that must match a JSON schema, use Extract data instead. |
Extract data (yekar.anthropic.extract-data) | Read | Ask a Claude model for an answer shaped by a JSON Schema, and get back the parsed object rather than prose to hand-parse. Use it to pull fields out of unstructured text, classify against a fixed set of labels, or produce a record another step consumes. Structured output is supported on the current models - if a model rejects it, List models will show which others are available. A refusal and a cut-off answer are the two ways to get no data, and both are reported explicitly. |
Count tokens (yekar.anthropic.count-tokens) | Read | Measure how many input tokens a prompt uses on a given Claude model, without generating anything or spending output tokens. Use it to size a prompt against a model's context window before sending it, to decide whether text needs splitting, or to estimate cost. Counts are model-specific - the same text tokenizes differently across model generations - so the answer is only valid for the model named. |
Notes
- In Anthropic, sign in to the workspace Yekar.AI should use - a key inherits that workspace's model access, rate limits and billing, so pick it deliberately.
- Go to platform.claude.com → API keys → Create key, and copy it (it starts with
sk-ant-). Anthropic shows it once. - Paste it here.
- Set a default model if you want one model used whenever a step names none; otherwise the tools fall back to a built-in default and every result reports which model actually ran.
- Call List models to confirm what this key reaches. Model ids carry no date suffix -
claude-sonnet-5, notclaude-sonnet-5-20260101- and a mistyped id reads as a 404. - Add one credential per Anthropic workspace - production and evaluation workspaces can live on the same connection.