Lifecycle: Experimental supporting surface. The canonical compiler and governed runtime live in
openadapt-flow. This tray is published on PyPI asopenadapt-tray, but it is a companion status surface, not a generally available integrated desktop product.
OpenAdapt is a governed demonstration compiler. You record a workflow once, it
compiles the demonstration into a deterministic program, and it replays that
program with zero model calls on the healthy path. When an interface drifts,
OpenAdapt re-resolves from retained evidence or proposes a governed repair, and
it halts instead of guessing when verification fails. Substrates are all
first-class in the product design (web, Windows, macOS, Linux, RDP, and
Citrix/VDI), with browser proven end to end today and the other substrates at
earlier maturity. That workflow logic belongs to openadapt-flow.
OpenAdapt Tray is a lightweight system-tray companion for the OpenAdapt desktop authoring experience. It does not record, compile, replay, repair, or train anything itself. It mirrors state and hands local actions to a companion desktop process over an authenticated loopback connection, and it reads a small needs-attention count from the hosted control plane.
From the tray you can:
- Start or stop recording on the desktop app, and see the live recording and compile lifecycle.
- Open recent captures from the local captures directory.
- See how many automations need attention and jump straight to them.
- Open the desktop app or the cloud dashboard.
- Pause or resume sync, shown as its own channel.
- Sign in (open the ingest-token settings page) and open settings.
The tray icon is always the OpenAdapt mark, tinted per lifecycle state so the state stays readable at a glance:
| State | Tint |
|---|---|
| Idle | Brand blue |
| Recording starting | Amber |
| Recording | Red |
| Recording stopping | Amber |
| Compiling | Purple |
| Error | Red |
The mark is stored once as a high-resolution transparent master and tinted from it, so every state renders as the same recognizable mark and the states never drift apart. Recording lifecycle and sync are modelled as two independent channels: a machine can be compiling a fresh recording while a previous push is still syncing, or sit idle while offline.
openadapt-tray (status mirror + launcher)
| \
| authenticated loopback IPC \ HTTPS: GET needs-attention count
v v
openadapt-desktop (local cockpit) hosted control plane (app.openadapt.ai)
|
v
openadapt-flow (compile / replay / halt / repair / teach)
- Local control goes to
openadapt-desktopover an authenticated loopback socket. The desktop app writes a discovery file the tray reads, then the tray sends start/stop, open-library, open-teach, and pause/resume-sync commands and consumes desktop status events. - Hosted status is a narrow read: the tray polls a needs-attention count and routes a click to the right place. It never uploads screenshots, bundles, or capture artifacts.
- The hosted-lifecycle behavior described here is merged on
mainand published to PyPI; the package classifier is Pre-Alpha. - Unit tests cover the client state machine, IPC framing, menus, the per-state icon tinting, and mocked hosted HTTP behavior. They do not prove a working desktop installer, a live hosted service, or an end-to-end authoring loop.
- The
openadapt-desktopmainbranch now serves the exact discovery-socket and command contract this tray expects, but the two surfaces have not been validated together end to end, and no signed, generally available desktop build ships that server yet. Treat the tray as a status surface until that integration is qualified.
The retired model-training controls and training states are not part of this release.
The menu is built from current local and hosted state:
Start Recording (<configured hotkey>)
Recent Captures
<N automations need attention> # only when count > 0
Open Desktop App
Open Cloud Dashboard
Pause Sync / Sync (offline)
Login...
Settings...
Quit
During a local operation, the recording item changes to Starting, Stop Recording, Stopping, or Compiling. These labels reflect events; the tray does not perform the work.
The tray discovers a local service from ~/.openadapt/desktop_ipc.json, then
uses an authenticated loopback connection. It can send commands to start or stop
recording, open the workflow library or teach surface, and pause or resume sync.
It also consumes desktop status events.
If discovery fails, the tray launches openadapt-desktop and waits about ten
seconds for the service. The desktop main branch now implements a matching
token-authenticated loopback server and writes this discovery file, but that
path has not yet been validated end to end from a shipped desktop build.
The poller calls:
GET <hosted_url>/api/needs-attention/count
Authorization: Bearer <ingest token>
The token is resolved from OPENADAPT_INGEST_TOKEN or the OS keychain and is
not written to tray.json. The default poll interval is 60 seconds, clamped to
at least 30 seconds, with a slower offline retry.
This is a narrow status endpoint, not hosted execution. The tray does not upload screenshots, workflow bundles, or capture artifacts through this poller.
cloud: a needs-attention click opens<hosted_url>/dashboard, which lists open halts and uncertain dispatches.byoc: while desktop IPC is connected, the click sendsopen_teachlocally so workflow data can remain in the customer environment.byocwithout desktop IPC: the current implementation falls back to the hosted dashboard. Regulated deployments must not treat this fallback as a validated PHI-safe path; it should be changed or policy-gated before production use.
pip install openadapt-tray
openadapt-tray # run the tray applicationTreat the install as an Experimental status surface, not a production desktop product (see the release boundary above).
git clone https://github.com/OpenAdaptAI/openadapt-tray.git
cd openadapt-tray
uv sync --extra dev
uv run pytest tests -q
uv run openadapt-trayRunning the process requires a graphical desktop/session. Without a compatible desktop IPC server or hosted token it may correctly remain offline, fail local actions, or open only browser routes.
For a runnable OpenAdapt workflow, use the canonical engine separately:
pip install openadapt-flow
openadapt-flow demo-record --out rec
openadapt-flow compile rec --out bundle --name my-task
openadapt-flow replay bundleNon-secret settings are stored at:
- macOS:
~/Library/Application Support/openadapt/tray.json - Windows:
%APPDATA%/openadapt/tray.json - Linux:
${XDG_CONFIG_HOME:-~/.config}/openadapt/tray.json
Representative settings:
{
"hotkeys": {
"toggle_recording": "<ctrl>+<shift>+r",
"open_dashboard": "<ctrl>+<shift>+d",
"stop_recording": "<ctrl>+<ctrl>+<ctrl>"
},
"captures_directory": "~/openadapt/captures",
"desktop_ipc_port": null,
"hosted_url": "https://app.openadapt.ai",
"deployment_lane": "cloud",
"poll_interval_s": 60,
"show_notifications": true,
"auto_start_on_login": false
}deployment_lane accepts cloud or byoc. The ingest token does not belong in
this file.
- The desktop socket contract exists on the desktop
mainbranch, but no shipped desktop build has been validated end to end with this tray. - No packaged installer proves tray startup, permissions, or auto-start across macOS, Windows, and Linux.
- Hosted polling is tested with mocks; repository tests do not validate a live service contract or service-level commitments.
- BYOC fallback can open the hosted dashboard when desktop IPC is absent.
- Recent-capture View still invokes a legacy launcher command before its file-browser fallback.
- Login opens an ingest-token settings page; the tray does not implement interactive authentication.
- The tray does not certify workflow safety or verify workflow effects.
src/openadapt_tray/
app.py tray lifecycle, desktop delegation, and routing
ipc.py authenticated loopback IPC client
hosted.py needs-attention poller and lane routing
state.py recording, sync, and badge state
menu.py state-dependent tray menu
icons.py per-state OpenAdapt-mark tinting
keychain.py ingest-token lookup
config.py non-secret local preferences
tests/ mocked/unit coverage for these client boundaries
| Project | Lifecycle and role |
|---|---|
openadapt-flow |
Canonical workflow compiler, runtime, certification, and governed repair engine |
openadapt-desktop |
Experimental authoring/teaching cockpit; its main now serves the IPC contract this tray expects, pending end-to-end validation |
OpenAdapt |
Flagship launcher and meta-repository |
Documentation for the wider stack lives at docs.openadapt.ai.
MIT. See LICENSE.