User repository on GitHub
Go SDK for TypeSafe AI.
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
Go client for the TypeSafe API with typed question and answer types, per-request model and timeout overrides, request-body passthrough for fields the SDK does not model, and slog diagnostics that exclude bodies and credentials.
User repository on GitHub
Go SDK for TypeSafe AI.
Repo stats from the GitHub API, cached Oct 2026.Homepage
This is one of several Jev listings for the same project. The canonical entry is TypeSafe SDK Go — Tangerg.
The module uses Go 1.27.1+. Install the SDK with: Set TYPESAFEAPIKEY before running the example. Create a client once and reuse it for later requests: Give each question a name in the Questions map. The response uses the same name for its answer. A request can combine all three question types:
Quoted from the project's README. Read the full README.
Watch out
The SDK encodes each request once and resends identical bytes on every attempt, so retrying after a connection failure can incur duplicate usage — the README says to disable retries if that is unacceptable for your application. The per-attempt timeout does not bound the overall retry budget; a context deadline is what limits total time, and a custom http.Client.Timeout adds a further independent limit. All HTTP redirects are rejected, including when you supply your own HTTPClient, so a gateway that answers with a redirect will simply fail. Debug logging records method, endpoint path, attempt number, status, duration and request ID, but a logged attempt record does not mean the response decoded successfully. Copy reviewed against the repository README on 2026-10-04, not from running it.
Original write-ups on what this entry is for.
Posts from builders shipping with Jev right now.