Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 21 additions & 2 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,16 @@ uvicorn llmgate.demo_app:app --port 8000
The local demo uses simulated providers. See examples.config.json and docs/DEPLOYMENT.md for configuration. Real upstreams require your own credentials; do not commit secrets. Ambiguous-failure failover is off by default; enabling it can duplicate generation or charges.

## Verification
The Actas 0.14.1 integration candidate passed 266 pytest cases, 61 end-to-end HTTP checks and a JavaScript smoke gate locally. The benchmarks use simulated providers and are not capacity claims. Optional real-provider tests skip without credentials.
The November 11, 2026 Redis-settlement maintenance candidate passed 274 pytest
cases, 61 local end-to-end HTTP checks, and JavaScript smoke checks in both
simulated-only and real-route-configured modes. The latter uses a local dummy
credential and does not call a real provider. Redis limiter tests execute the
actual Lua 5.1 scripts rather than reimplement their arithmetic.

These are dated local results, not a claim that every future commit passes.
See [GitHub Actions](https://github.com/Rajatendu1/llmgate/actions) for hosted
results at a particular head. The benchmarks use simulated providers and are
not capacity claims. Optional real-provider tests skip without credentials.

## How llmgate uses Actas

Expand Down Expand Up @@ -41,4 +50,14 @@ https://llmgate-demo.vercel.app/
The public demo labels simulated and real-model paths. Redis-backed limits are optional; breaker and latency state are per process. Usage telemetry is not billing-grade. The Docker packaging has not been built or verified.

## Repository setup
This initial publication does not connect GitHub to Vercel or enable GitHub Actions. The existing CI definition is retained as docs/ci.example.yml, outside the active workflow directory. No license has been added to this repository.
GitHub is the source of truth for project changes. The active
[workflow](.github/workflows/ci.yml) runs pytest, local HTTP end-to-end tests
and a local JavaScript smoke test for pull requests and pushes to main.
Protected main requires its source check before normal merge. Vercel is
connected to this repository; preview builds are distinct from production.
Verify a staged production build before promotion, then check live workflows
end to end, including the changed behavior and failure cases. A ready build
or health probe alone is not proof of token settlement or real-provider access.

The historical example at `docs/ci.example.yml` is not the active workflow.
No license has been added to this repository.
29 changes: 28 additions & 1 deletion docs/DEPLOYMENT.md
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,34 @@ attempt counts, and no prompt content or key material), and the real-model and
backup-model Try paths failing closed with a clean 404 when no real route is
configured.

The run is mirrored to `e2e/E2E_TRANSCRIPT.txt`. The committed transcript records a historical local run. A sample CI definition is retained at `docs/ci.example.yml`; it is not active. Run `pytest`, `bash e2e/run_e2e.sh` and `bash e2e/js_smoke.sh` locally.
The run is mirrored to `e2e/E2E_TRANSCRIPT.txt`. The committed transcript records
a historical local run. The active workflow in `.github/workflows/` runs pytest,
this HTTP suite and the local JavaScript smoke test. `docs/ci.example.yml` is
only a historical example. Run the same gates locally with:

```bash
python -m pytest -q
bash e2e/run_e2e.sh
bash e2e/js_smoke.sh
bash e2e/js_smoke.sh --simulated-only
```

## GitHub-backed deployment checks

Use a pull request and the required source check before normal merge to main.
Do not bypass branch protection. Vercel builds from the repository; preview
and production environments can have different provider and store settings.
Check the exact source head, staged production runtime and configuration before
promotion. After promotion, check the live workflows end to end, including the
changed behavior and relevant failures, and inspect desktop/mobile layouts.
Keep the previous deployment available for rollback.

A ready build or `/healthz` response alone does not establish token settlement,
upstream model availability or pricing. For a settlement change, use a bounded
simulated-only key to check two large token reservations after small actual
responses. Confirm the second admission occurs too soon for ordinary refill
to explain it. If the result is inconclusive, say so. Do not call real providers
or spend money just to turn an untested path into a passing check.

## Environment knobs

Expand Down
Loading