Overview
The HAR-derived API turns browser HAR archives into structured, parameterized API client specs. Upload a HAR file captured from a web session, and the endpoint extracts repeatable API calls, templatizes path, query, and body parameters, and emits code-ready client templates in Python, TypeScript, or curl. All routes are mounted under/api/har-derived-api and require a valid bearer token via the platform auth middleware.
Base URL
Authentication
All requests must include a valid bearer token in theAuthorization header. See Authentication for details.
Endpoints
Ingest a HAR archive
log.entries array, and returns the API calls that look repeatable. Static assets, error responses, and duplicate requests are automatically filtered out.
Request body
Example request
Example response
Endpoint object
TemplatedParam object
IngestStats object
Generate a client
python, typescript, and curl.
Request body
Example request
Example response
Filtering and parameter inference
The ingest route applies the following heuristics:- Entries with a response status
>= 400are skipped. - Static asset paths ending in
.js,.css,.png,.jpg,.jpeg,.gif,.svg,.ico,.woff,.woff2,.ttf,.eot,.otf,.map,.json,.xml, or.webpare skipped. - Duplicate
METHOD URLpairs are deduplicated, keeping the first occurrence. - Hop-by-hop and sensitive headers such as
Authorization,Cookie,X-Api-Key, andApi-Keyare stripped. - Path segments that look like integers, UUIDs, or long hyphenated strings are converted to
{id},{id2}, and so on. - Query and JSON body values longer than 8 characters, numeric values, or strings that look like secrets are marked as templated.