Skip to content
Demo

settings.app

ANTHROPIC CLAUDE

Add your Claude key to generate questions

Create a key in the Anthropic console
Where does my key go?

The key is only sent to api.anthropic.com, inside a header of each request, over an HTTPS connection. It stays in memory by default. Like any request made from your browser, it is visible in your developer tools.

Claude model

Describe. Adjust. Copy.

Jev makes the call in a split second. EasyJev helps you talk to it in your own words.

Describe.

Describe what Jev should decide: Claude writes the questions for you.

describe.app
0 / 4000 characters (min. 10)
Jev will read:
`message`
optional

In n8n, open the node right before Jev, switch its OUTPUT panel to JSON view, copy everything and paste here. What you paste stays in your browser: only the field names are sent to Claude, never the values.

Replace current content?

Your editor already contains work (modified questions or fields). Continuing will replace current fields and questions.

No Claude key? Start from a template:

Replace current content?

Your editor already contains work (modified questions or fields). Continuing will replace current fields and questions.

Adjust.

Check what Jev reads and tweak the questions.

adjust.app

What Jev reads

This is the field from the previous node that Jev analyzes. Click its name to rename it.

Questions for Jev

Add your first question

Pick the format that best fits your decision:

Copy.

Copy the node, paste it into your n8n workflow: the request is already set up.

copy.app
Errors to fix
Your Jev access

Fix blocking errors below to copy the n8n node.

Then, in n8n

  1. 1. Paste the node

    Click on your n8n workflow canvas, then press Ctrl+V (Cmd+V on Mac). A "Jev" node appears with all its settings. Connect it to the node providing the text to analyze.

  2. 2. Add your key (once)

    Open the node; under Credential for Header Auth, choose Create new credential; set Name = Authorization and Value = Bearer (with a space) followed by your TypeSafe key (from console.typesafe.ai), then click Save. Note: your key stays in n8n’s saved credentials, never in EasyJev nor in the copied node.

  3. 3. Check

    In the JSON field, the fx icon is active and the preview shows your text between quotes (the node reads the field from the preceding node: make sure the field name exists in its input data).

  4. 4. Read answers

    Each question outputs a result ready to use in subsequent nodes (If, Switch, etc.):

    Configure valid questions in the editor to display tailored reading expressions.

n8n-help.app

INTEGRATION GUIDE
Configure manually (if paste does not work)

  1. Block 1 — The node

    1. Add an HTTP Request node

    In your n8n workflow, insert an HTTP Request node. This is the node that sends the request to Jev.

  2. 2. POST Method

    In Method, choose POST to transmit the data to evaluate.

  3. 3. Jev URL

    In URL, paste the TypeSafe API address: https://api.typesafe.ai/v1/systemone (or click Copy URL above).

  4. Block 2 — The key (authentication)

    4. Generic authentication

    In Authentication, choose Generic Credential Type.

  5. 5. Header Auth credential type

    In Generic Auth Type, choose Header Auth to transmit your key in an HTTP header.

  6. 6. Create and save credentials

    Create the credential: Name = Authorization, Value = Bearer followed by your TypeSafe key (from console.typesafe.ai), then click Save. Your key stays stored in n8n and is never sent to EasyJev.

  7. Block 3 — The header declaring JSON

    7. Enable headers

    Toggle on Send Headers. A header is metadata describing the transmission to the API.

  8. 8. Specify headers below

    In Specify Headers, choose Using Fields Below to define the header manually.

  9. 9. Add Content-Type header

    Under Headers, add an entry with Name = Content-Type and Value = application/json. This informs Jev that the sent content is formatted as JSON.

  10. Block 4 — Sent content

    10. Enable request body

    Toggle on Send Body. The body is the main payload sent to Jev (the fields to analyze and your questions).

  11. 11. JSON content type

    In Body Content Type, choose JSON.

  12. 12. Specify body as JSON

    In Specify Body, choose Using JSON to paste the JSON payload generated by EasyJev directly.

  13. 13. Paste JSON in Expression mode

    In JSON, paste the JSON copied from EasyJev. Upon pasting, n8n automatically switches the field to Expression mode (fx icon on the left, verified on n8n 2.33.4). An n8n expression (wrapped in double braces {{ ... }}) dynamically computes the value at runtime. Check the preview below the field: it should display your text between quotes. If the field stays in plain text, click the fx icon to switch to Expression manually. With multiple fields, each gets its own expression (e.g. {{ JSON.stringify($json.message) }}).

  14. Block 5 — Read answers

    14. Read answers

    Each question outputs a result ready to use in subsequent nodes (If, Switch, etc.):

Troubleshooting

  • If pasting does nothing or n8n reports an unknown node: expand the "Configure manually" section above to set up the node step by step.
  • Unreadable or empty body: verify the Content-Type header (steps 7 to 9) and that the JSON field is in Expression mode (step 13).
  • 401: Missing, mistyped, or expired API key.
  • 422: Malformed request body (missing model or scale outside 2-10).
  • 429 / 529: Rate limit reached or temporary service overload.
Quick key check:
curl -H "Authorization: Bearer <KEY>" https://api.typesafe.ai/v1/models

Reset everything?

Your fields, questions and current description will be cleared. Your Claude key and model are not affected.