Cloud scheduling backend for Lamia. Deploy .lm scripts to run on a schedule in the cloud — fully automated. Currently supports GCP.
pip install "lamia-lang[cloud]"- GCP project with billing enabled
- Application Default Credentials:
gcloud auth application-default login
All required GCP APIs (including Service Usage) are enabled automatically on first deploy.
- Add a
cloudsection to your project'sconfig.yaml:
cloud:
provider: gcp
project_id: my-gcp-project
location: us-central1 # optional, default: us-central1- Schedule your script with the
--remoteflag:
lamia schedule add my_script.lm --every day --remoteThe --remote flag tells lamia to deploy and run the script in the cloud instead of locally.
lamia schedule list # shows all jobs (local + cloud) with live status
lamia schedule add X --remote # deploy and schedule a new cloud job
lamia schedule remove <id> # tears down cloud resources and removes the joblamia schedule add --remotepackages your project and deploys it as a Cloud Run service- Cloud Scheduler triggers it on your cron schedule
- Logs are available in Cloud Logging
lamia schedule listfetches live execution status from the cloud
Scripts that use LLM calls run through Vertex AI on cloud. This gives you:
- No API keys — authentication via IAM, no keys to store, rotate, or leak
- Budget control — Vertex AI quotas and billing alerts
- Secure by default — no API key transport or storage, traffic stays within GCP
| Provider | Cloud routing |
|---|---|
| Anthropic (Claude) | Runs natively on Vertex AI — same models, same quality |
| Google (Gemini) | Runs natively on Vertex AI |
| OpenAI (GPT, o-series) | Automatically mapped to Gemini by tier (strong/medium/light) with runtime selection of the best available current Gemini model |
Anthropic and Google models run as-is. OpenAI models are mapped because they're not available on Vertex AI — tier classification is stable while the selected Gemini model is discovered dynamically at runtime.
| Field | Required | Default | Description |
|---|---|---|---|
cloud.provider |
Yes | — | Cloud provider (currently gcp) |
cloud.project_id |
Yes | — | Your GCP project ID |
cloud.location |
No | us-central1 |
Region for Cloud Run deployment |
No environment variables are required.
- If Vertex AI access is not enabled yet, lamia-cloud logs a project-specific URL and attempts to open it automatically in your browser:
https://console.cloud.google.com/vertex-ai?project=<your-project-id> - After accepting terms, re-run the schedule/install command once.
git clone https://github.com/lamia-lang/lamia-cloud.git
cd lamia-cloud
pip install -e ".[dev]"
pytestgit tag v0.1.0
git push origin v0.1.0