devx-geodeploy
Pulumi example where Jev picks the best cloud region and modules per user location while code does pricing math and provisioning.
Overview
GeoDeploy works out where a container should run and what it will cost, then deploys it. Given users' locations and a container image, it picks the best AWS, Azure or Google Cloud region for each location with Jev and provisions consistent Kubernetes environments with the Pulumi Automation API.
It illustrates an agentic infrastructure workflow where each system has one job: the TypeScript app takes the request, fetches live pricing and does all arithmetic; Jev decides which region wins and which modules to use, and never does maths, generates code or invents region names; Pulumi provisions clusters, registries, workloads and DNS and decides nothing. Every Jev answer is checked against a confidence threshold, and a deterministic fallback is used if confidence is low, the API is unreachable or no key is set. The plan command touches no cloud account and writes plan.json and plan.md.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which cloud region should serve users in this location?(inferred) | — |
| Choice | Which infrastructure modules should this deployment use?(inferred) | — |
How it works
- Read user locations and container image from the request
- Fetch live pricing and rank candidate regions in code
- Jev chooses the winning region and modules; low confidence uses the fallback
- Pulumi provisions the environments
Strengths
- Clear separation of judgment, arithmetic and provisioning
- Deterministic fallback without a key
- Plan step needs no cloud credentials
Caveats
- Example project rather than a product
- Cloud credentials and Pulumi token needed to apply
Getting started
pnpm install, pnpm build, cp .env.example .env, then node packages/cli/dist/index.js plan --request examples/sample-request.json --out ./out.At a glance
- Type
- Project
- Status
- prototype
- Language
- TypeScript
- Runtime
- Node CLI + Pulumi
- Has code
- yes
- License
- Apache-2.0
Use cases and integrations
How curated lists describe it
Cost-optimized, latency-aware multi-cloud Kubernetes placement with Pulumi and TypeSafe Jev.
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
Which region wins for each user location and which modules to use.
Does Jev compute prices?
No, the TypeScript app does all arithmetic; Jev only makes judgment calls.
Can I run it without a TypeSafe key?
Yes, it falls back to a deterministic choice.
Related projects
OpenClaw plugin and CLI that turn a plain-language condition into a scheduled trigger, waking the model only when Jev says it is true.
PluginbetaTypeScriptGo event bus whose optional routing layer asks Jev a Choice to pick an event projection only for ambiguous cache misses.
ProjectGoReusable project context and specialist workflows for existing AI tools, with an addon that offloads bounded decisions to Jev.
ProjectproductionTypeScriptVerifiable coding-agent runtime built on checkable execution graphs, listed with an optional Jev provider for bounded choices like routing.
ServiceproductionTypeScriptAutoGPT agent platform, listed as adding Jev decision blocks for typed routing, filtering, scoring and confidence-gated next actions.
ProjectproductionPythonMulti-agent orchestration layer for Claude Code with opt-in Jev hooks for fast judgment points, decision caching and egress controls.
PluginproductionTypeScript