@conclave-ai/platform-render

Conclave AI Render adapter. Resolves a deploy URL for a given commit SHA via Render's REST API (api.render.com/v1).

Ranked #1 deploy target for solo makers outside the original v2.0 five (Vercel/Netlify/Cloudflare/Railway/deployment-status) per the 2026-04 adapter scope study.

Install

Pulled in automatically by @conclave-ai/cli when "render" is in config.visual.platforms.

pnpm add @conclave-ai/platform-render

Env

| Var | Required | Notes | |---|---|---| | RENDER_API_TOKEN | yes | User-scoped token (or Service Preview PAT) | | RENDER_SERVICE_ID | yes | srv-xxxxxxxxxxxx — either the main service or a specific Service Preview service |

Behavior

• GET /v1/services/{serviceId} — resolve canonical URL.
• GET /v1/services/{serviceId}/deploys?limit=20 — list deploys.
• Client-side filter: deploy.commit.id === sha AND deploy.status === "live".
• Newest by finishedAt (fallback createdAt) wins.
• Return { url: service.serviceDetails.url, sha, deploymentId, createdAt }.

Caveats

• No per-deploy preview URL on standard Web Services. The service URL is stable; returning it + the matched deploy ID is the best conforming signal for the Platform interface. If you need per-PR URLs, configure Render's Service Previews — each PR spawns its own service with its own srv-... id; point this adapter at the preview service (or register multiple adapter instances).
• 404 on service → adapter returns null (not thrown). 401/403 → throws. 5xx → throws with truncated body for diagnostics.
• No native "wait for SHA" API — waitSeconds polls every ~3s until deadline.