feat(deploy): one-click cloud deployment for OpenHuman Core (closes #1280)#1304
Merged
senamakel merged 3 commits intotinyhumansai:mainfrom May 6, 2026
Merged
Conversation
Closes tinyhumansai#1280 — packages the headless Rust core for cloud hosting. - docker-compose.yml at repo root: single-service compose wrapping the existing Dockerfile, named volume for OPENHUMAN_WORKSPACE, env-file pattern, healthcheck, restart unless-stopped. - .do/app.yaml: DigitalOcean App Platform spec referencing the Dockerfile, /health probe, secret-flagged OPENHUMAN_CORE_TOKEN. Enables the "Deploy to DigitalOcean" button in docs. - docs/CLOUD_DEPLOY.md: cloud deploy guide covering DO App Platform (one-click + doctl) and any-VPS Compose flow. Documents required envs, ports, persistence, health, restart, update path, and how to point the desktop app at a hosted core. - .github/workflows/deploy-smoke.yml: PR-gated smoke that builds the image, boots the container, polls /health, and confirms /rpc rejects unauthenticated calls and accepts the configured bearer token. Triggers on changes to deploy artifacts and src/**. - README.md: cross-link to the cloud deploy doc. The Dockerfile and OPENHUMAN_CORE_* env knobs already existed; this PR bundles them into a one-click flow with regression coverage.
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (2)
✅ Files skipped from review due to trivial changes (1)
📝 WalkthroughWalkthroughAdds cloud deployment artifacts and docs: a DigitalOcean App Platform manifest, Docker Compose file, expanded Docker runtime/build deps, a GitHub Actions smoke-test workflow that builds and validates the container, and a CLOUD_DEPLOY guide with a README link. No public API/signature changes. ChangesCloud Deployment Infrastructure
Sequence DiagramsequenceDiagram
actor Trigger as GitHub Push/PR/Manual
participant GHA as GitHub Actions
participant Builder as Buildx/Docker
participant Container as Running Container
participant Health as /health Endpoint
participant RPC as /rpc Endpoint
Trigger->>GHA: start deploy-smoke workflow
GHA->>Builder: checkout & build image
Builder-->>GHA: image built and loaded
GHA->>Container: run container with env & ports
Container->>Health: initialize and listen on port
loop Health Check (up to 30 retries)
GHA->>Health: GET /health
Health-->>GHA: 200 OK or retry
end
GHA->>RPC: GET /rpc (no Authorization)
RPC-->>GHA: 401 Unauthorized (expected)
GHA->>RPC: GET /rpc (with Bearer token)
RPC-->>GHA: 200 OK (expected)
GHA->>Container: dump logs
GHA->>Container: stop & remove container
GHA->>Trigger: report results
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (2)
.do/app.yaml (1)
51-56: 💤 Low valuePlaceholder token is appropriately marked SECRET, but consider adding a validation hint.
The placeholder
CHANGE_ME_BEFORE_DEPLOYis intentional per the PR objectives, andtype: SECRETensures App Platform treats it securely. However, operators who forget to change it will have a guessable token.Consider adding a comment indicating that App Platform will reject this specific placeholder, or update the deploy docs to emphasize the token must be changed before the app becomes reachable.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.do/app.yaml around lines 51 - 56, Add a validation hint for the OPENHUMAN_CORE_TOKEN secret so operators know the placeholder "CHANGE_ME_BEFORE_DEPLOY" must be replaced; update the .do/app.yaml entry for OPENHUMAN_CORE_TOKEN to include a clear comment immediately above or beside the value that App Platform will reject this placeholder (or reference the deploy docs), and/or add a short note in the deployment documentation stating that OPENHUMAN_CORE_TOKEN must be changed from "CHANGE_ME_BEFORE_DEPLOY" to a strong random secret before deployment..github/workflows/deploy-smoke.yml (1)
61-69: 💤 Low valueConsider adding
OPENHUMAN_CORE_HOSTto match production configuration.The container is started without
OPENHUMAN_CORE_HOST=0.0.0.0. While this likely works if the Dockerfile sets a default, explicitly setting it would ensure the smoke test mirrors the production environment defined indocker-compose.ymland.do/app.yaml.🔧 Optional: align env vars with production specs
docker run -d \ --name oh-smoke \ -p 7788:7788 \ -e OPENHUMAN_CORE_TOKEN=ci-smoke-token \ + -e OPENHUMAN_CORE_HOST=0.0.0.0 \ -e OPENHUMAN_APP_ENV=staging \ -e BACKEND_URL=https://staging-api.tinyhumans.ai \ openhuman-core:smoke🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.github/workflows/deploy-smoke.yml around lines 61 - 69, Add the missing OPENHUMAN_CORE_HOST env var to the docker run invocation that starts the smoke container (the step that creates container name oh-smoke) so the smoke test mirrors production; update the Run container step to include -e OPENHUMAN_CORE_HOST=0.0.0.0 alongside the existing OPENHUMAN_CORE_TOKEN, OPENHUMAN_APP_ENV and BACKEND_URL environment variables to ensure the container binds the same host as in docker-compose.yml/.do/app.yaml.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/CLOUD_DEPLOY.md`:
- Around line 13-15: The three table-of-contents links using "DigitalOcean App
Platform — one-click", "DigitalOcean App Platform — manual `doctl`", and "Any
VPS via Docker Compose" use em-dashes that don't match GitHub's generated
heading anchors; update either the headings or the fragment links so they match
exactly — simplest fix: replace the em-dash (—) with a regular hyphen (-) in the
headings or change the link fragments to the exact anchors GitHub generates
(e.g., "digitalocean-app-platform-one-click" and
"digitalocean-app-platform-manual-doctl") so the fragment identifiers resolve
and the markdown link checker passes.
---
Nitpick comments:
In @.do/app.yaml:
- Around line 51-56: Add a validation hint for the OPENHUMAN_CORE_TOKEN secret
so operators know the placeholder "CHANGE_ME_BEFORE_DEPLOY" must be replaced;
update the .do/app.yaml entry for OPENHUMAN_CORE_TOKEN to include a clear
comment immediately above or beside the value that App Platform will reject this
placeholder (or reference the deploy docs), and/or add a short note in the
deployment documentation stating that OPENHUMAN_CORE_TOKEN must be changed from
"CHANGE_ME_BEFORE_DEPLOY" to a strong random secret before deployment.
In @.github/workflows/deploy-smoke.yml:
- Around line 61-69: Add the missing OPENHUMAN_CORE_HOST env var to the docker
run invocation that starts the smoke container (the step that creates container
name oh-smoke) so the smoke test mirrors production; update the Run container
step to include -e OPENHUMAN_CORE_HOST=0.0.0.0 alongside the existing
OPENHUMAN_CORE_TOKEN, OPENHUMAN_APP_ENV and BACKEND_URL environment variables to
ensure the container binds the same host as in docker-compose.yml/.do/app.yaml.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 6912b7dc-04c9-427c-9074-4ca679479d7b
📒 Files selected for processing (5)
.do/app.yaml.github/workflows/deploy-smoke.ymlREADME.mddocker-compose.ymldocs/CLOUD_DEPLOY.md
Two CI fixes for tinyhumansai#1304 (tinyhumansai#1280): - Dockerfile: add libasound2-dev / libxdo-dev / libxtst-dev / libx11-dev / libevdev-dev to the builder stage and the matching libasound2 / libxdo3 / libxtst6 / libx11-6 / libevdev2 to the runtime stage. The core crate has unconditional dependencies on cpal, enigo, arboard, and rdev (voice / autocomplete / clipboard subsystems) that link against ALSA and X11, even when those features are inactive at runtime. Without these packages the bookworm build of openhuman-core fails at the alsa-sys pkg-config step. - docs/CLOUD_DEPLOY.md: replace em-dashes in section headings with colons so GitHub's slug algorithm produces single-hyphen anchors matching the TOC links. Lychee was failing on 'tinyhumansai#1-digitalocean-app-platform-one-click' / '...manual-doctl' because the actual headings produced double-hyphen slugs.
senamakel
approved these changes
May 6, 2026
This was referenced May 7, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Packages the headless Rust core (
openhuman-core) for one-click cloud hosting. The Dockerfile and bind-address env knobs already existed; this PR bundles them into deployable artifacts plus regression coverage.docker-compose.yml(new, repo root) — single-service Compose wrapping the existing Dockerfile, named volume forOPENHUMAN_WORKSPACE, env-file pattern, healthcheck,restart: unless-stopped. Validated withdocker compose config -q..do/app.yaml(new) — DigitalOcean App Platform spec referencing the same Dockerfile,/healthprobe, andOPENHUMAN_CORE_TOKENflagged as aSECRET. Enables the "Deploy to DO" button in the docs. Validated withdoctl apps spec validate --schema-only.docs/CLOUD_DEPLOY.md(new) — three-path deploy guide: DO one-click, DO viadoctl, any-VPS Compose. Documents required envs, ports, persistence, health, restart, update path, and how to point the desktop app at a hosted core..github/workflows/deploy-smoke.yml(new) — PR-gated smoke: build image, boot container, poll/health, assert/rpcreturns 401 without bearer and 200 with the configuredOPENHUMAN_CORE_TOKEN. Triggers on changes to deploy artifacts andsrc/**. Satisfies the regression-safety bullet in the issue.Dockerfile— added Linux ALSA + X11 + input system dependencies (libasound2-dev,libxdo-dev,libxtst-dev,libx11-dev,libevdev-devto the builder; matching runtime libs to the runtime stage). Required becausecpal/enigo/arboard/rdevare unconditional crate deps; without them the bookworm build ofopenhuman-corefails atalsa-sys.README.md— cross-link to the new deploy doc.Problem
Per #1280: OpenHuman has no documented one-click cloud deploy. Users / internal testers who want a hosted core (instead of a local sidecar) have nothing to point at — there's a Dockerfile but no Compose, no provider template, no docs, and no smoke coverage to keep the deploy path from silently breaking. The Dockerfile itself was also missing Linux audio/X11 build deps and would fail to compile on bookworm.
Solution
Ship the smallest credible deploy bundle:
.do/app.yamlso the "Deploy to DigitalOcean" button works straight from the docs.docker-compose.ymlfor any VPS, with persistent named volume (App Platform Basic doesn't have block storage; this is the durable path).docs/CLOUD_DEPLOY.mdcovers both flows, required envs, the bearer-token contract, and the "point my desktop app at a hosted core" handoff.deploy-smoke.ymlbuilds the image, boots it, and probes/health+/rpc(401 without bearer, 200 with) on every PR that touches deploy artifacts orsrc/**.No Rust or TS code changed; the existing core already binds
0.0.0.0:7788and enforces bearer auth on/rpc.Submission Checklist
docs/TESTING-STRATEGY.md—deploy-smoke.ymlcovers the happy path (/health200, authed/rpc200) and the failure path (unauthed/rpc401).docs/TEST-COVERAGE-MATRIX.md.## Related— N/A, see above.localhost; no new mock backend changes.Closes #NNNin the## Relatedsection — see below.Impact
basic-xsinstance (~1 vCPU / 512 MB).OPENHUMAN_CORE_TOKENis the gate to/rpc; the spec marks ittype: SECRETand the doc flags the placeholder swap as a hard prerequisite./health,/events,/ws/dictationremain public persrc/core/auth.rs:49.Test plan
docker compose config -qparsesdocker-compose.yml.doctl apps spec validate .do/app.yaml --schema-onlyaccepts the App Platform spec..github/workflows/deploy-smoke.yml./rpcprobe uses a real registered method (openhuman.about_app_list— exercised bytests/json_rpc_e2e.rs:2656).cargo build --releaseopenhuman-corewithOPENHUMAN_CORE_TOKEN=test-token:GET /health→200 {"ok":true}POST /rpcno bearer →401POST /rpccorrect bearer →200(returns 83 capabilities)POST /rpcwrong bearer →401deploy-smoke.ymlis wired up and runs on this PR; the green status here IS the test (full Docker image build + container boot + same/healthand/rpcprobes against the running container)..do/app.yamlreferencestinyhumansai/openhuman:main, so the "Deploy to DO" button only resolves once this PR has landed). Pre-merge validation is covered by schema-validated spec + green CI image build.Notes for reviewers
OPENHUMAN_CORE_TOKEN: CHANGE_ME_BEFORE_DEPLOYin.do/app.yamlis intentional — the docs call out replacing it withopenssl rand -hex 32in the App Platform UI before the first deploy. The field istype: SECRETso it is encrypted at rest.basic-xsdoes not provide block storage; the workspace lives on the container's ephemeral FS and is wiped on redeploy. The doc points operators at the Compose path on a Droplet for durable storage and flags this explicitly under the App Platform section.Dockerfileare technically a fix for a pre-existing image-build bug (the file existed but had never been built in CI). Bundling them here keeps the deploy-smoke workflow as the regression gate that would have caught this earlier.Related
Closes #1280
Summary by CodeRabbit
New Features
Documentation