AI Decision System
Replayable, offline-verifiable records for AI-assisted go/no-go decisions: experiment promotions, typed model answers and human-approved actions.
Publish OpenAPI specs, price per call, and let humans and agents pay with prepaid credits.
Developer toolsTypeScript · JavaScriptGitHub @zevium-devOnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Run /nomorevibe verify in your project folder to verify ownership and manage the introduction yourself.
Verify in one minuteAsk about other verification options
Zevium
Translation pending
Taken from the repository README
Loading the full README. The saved excerpt is shown below.
Zevium lets developers and agents call APIs with one key and a prepaid organization wallet. Publishers list their APIs with an OpenAPI spec, set prices per endpoint, and earn 95% of each successful paid call. Consumers can try free mock responses before making live calls.Read the full README ↗Zevium lets developers and agents call APIs with one key and a prepaid organization wallet. Publishers list their APIs with an OpenAPI spec, set prices per endpoint, and earn 95% of each successful paid call. Consumers can try free mock responses before making live calls.
x-zevium-cost, x-zevium-free-tier) live in the spec. Published specs cannot be edited. New versions replace the spec used by the gateway./gateway/{org}/{project}/{path} with one API key. Agents get an
MCP endpoint and a machine-readable /discovery index./mock mode, /mcp + /discovery
for agents./docs.apps/web/ # TanStack Start app — all screens
apps/gateway/ # CF Worker: metered proxy, wallet DO, mock, MCP, discovery
convex/ # Convex schema + functions (control plane)
packages/shared/ # OpenAPI parsing, x-zevium-* pricing, validation
e2e/ # agent-browser end-to-end suite (auth, publisher, consumer)
.agents/notes/ # committed agent memory: product, features, architecture, decisions, research
Docs: everything lives in .agents/notes/ — start with the product overview, roadmap, features, and architecture.
Security policy: SECURITY.md
pnpm install
pnpm exec convex dev # control plane (needs CONVEX_DEPLOYMENT in .env.local)
pnpm dev # web app on :3000
cd apps/gateway && pnpm exec wrangler dev --port 8787 # gateway
pnpm seed # test user + org ([email protected])
Verification:
pnpm ci:pr # formatting, lint, static checks, types, tests, and build
bash e2e/run-all.sh # browser e2e: auth, publish, consume
pnpm build is the canonical root production build. Supply
CLERK_SECRET_KEY, VITE_CLERK_PUBLISHABLE_KEY, VITE_CONVEX_URL, and
VITE_GATEWAY_URL in the environment, or place them in the ignored
apps/web/.env.production.local. The wrapper derives the full Git SHA,
forwards only the four public build values through Turbo's hashed environment,
and passes the Clerk secret through without caching it. Production output must
never contain dist/server/.dev.vars.