Files
opencode-mobile/docs/marketing/show-hn.md
engineer e39cec3847 fix(launch): Expo SDK 52->54 in HN/Reddit copy; add reply-flow E2E evidence; version refs
- Launch posts claimed Expo SDK 52 but app is on SDK 54 (factual accuracy
  before public posting — HN/Reddit devs check this)
- docs/qa/REPLY-FLOW-E2E-2026-06-08.md: verified send->streaming reply works
  against the live opencode server via app-identical sdk.ts calls (free model);
  closes the 'opencode can't reply in CI' residual at the data-contract level
- owner-submissions.md: 0.4.3 -> 0.4.4

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-08 07:42:19 -07:00

3.7 KiB
Raw Blame History

Show HN — copy-paste ready

Post once. HN allows a single submission per project. Tue–Thu, 8:00–10:00am US Eastern is the sweet spot. {{PLAY_URL}} = swap in https://play.google.com/store/apps/details?id=cc.agentlabs.opencode on the day Play approval lands. Until then, leave the Play line exactly as written below.


Title (use exactly — under 80 chars, no emoji, no "free"):

Show HN: OpenCode Mobile – drive your self-hosted AI coding agent from your phone

URL field: https://github.com/dzianisv/opencode-mobile


Text (paste into the "text" box if you submit as a text post; otherwise post as the first comment immediately after submitting):

Hey HN,

I built a React Native / Expo Android app that lets you control a running opencode session from your phone.

What it does

opencode (sst/opencode, MIT) is a terminal-based AI coding agent — you run it on your laptop or a server, it talks to Claude / GPT / Gemini / local models via your own API keys, and it writes, edits, and runs code through tool calls. OpenCode Mobile is a thin HTTP + SSE client that connects to that server process. From your phone you can:

  • Watch token-by-token streaming output as the agent reasons and codes
  • Review inline file diffs before they're written
  • Approve or reject tool calls (file writes, shell commands) in real time
  • Start, resume, and switch between sessions
  • Connect over local Wi-Fi, Tailscale, Cloudflare Tunnel, or ngrok

Why I built it

I run long opencode sessions that outlast my desk time. I wanted to check in from the couch, approve a tool call, and let the agent keep going. The alternatives — SSH + a terminal app, or generic Claude/GPT chat apps — lose the structured session state: no diffs, no tool-call gating. The opencode HTTP API was already there; the phone just needed a real client.

How it's built

  • React Native + Expo SDK 54, TypeScript throughout
  • SSE streaming via an EventSource polyfill (RN has no native one); needed patching because Android's OkHttp drops the Accept: text/event-stream header across redirects, which breaks reconnects
  • Diff rendering with a custom line-level parser into a FlatList — no WebView, so dark-mode theming stays consistent and it's cheap to render
  • Biometric auth via expo-local-authentication gates app-open and individual sends
  • Connection secrets in the Android Keystore via expo-secure-store
  • CI builds a signed AAB on every tag via GitHub Actions + EAS

What it is not

Not a standalone AI model and not a code editor. You need a running opencode server. If you already use opencode, setup is:

npm install -g opencode-ai
OPENCODE_SERVER_PASSWORD=yourpassword opencode serve --hostname 0.0.0.0 --port 4096

…then paste the URL into the app. Your API keys stay on your server; nothing proxies through any backend of mine.

Status / privacy

Stable on Android, MIT licensed. No accounts, no ads. Sentry crash reporting is opt-in and off by default. Install today via a self-hosted F-Droid repo or a direct signed APK. Google Play is in review and should be live shortly — {{PLAY_URL}} (currently in review; until then use the links below).

Happy to answer questions about the SSE streaming implementation, the Expo build pipeline, or the opencode API surface.