Orange3 Canvas Agent
Orange3-CanvasAgent adds a new Orange widget named Canvas Agent. It lets
you describe a workflow in plain language, previews the agent's structured
canvas changes, and keeps those changes reversible until you choose
Keep Changes or Revert AI Commit.
Version 0.1.3 focuses on install polish: a simpler Windows installer flow,
clearer first-run guidance, and better repair steps when the widget is installed
but not visible yet.
Release Links
- GitHub releases: https://github.com/pantagram1031/orange3-canvasagent/releases
- PyPI package: https://pypi.org/project/Orange3-CanvasAgent/
- Issue tracker: https://github.com/pantagram1031/orange3-canvasagent/issues
Install For Beginners On Windows
This is the recommended path if you already use the normal Orange desktop app and do not want to open a terminal.
- Open the GitHub releases page.
- Download
CanvasAgentSetup.exefrom the latest release. - Double-click the installer.
- If Windows shows a security warning, verify the file came from the GitHub release page above, then continue only if you trust it.
- In the installer window, wait for Orange detection to finish.
- Select your Orange installation from the list.
- If nothing is listed, click Choose Folder... and point to your Orange install folder.
- Click Install / Repair.
- Continue to Verify and run the smoke check.
- On Finish, confirm the success message and click Open Orange.
- When Orange launches, look in the widget toolbox for the Canvas Agent category.
- Drag Canvas Agent onto the canvas.
If the widget still does not appear, jump to Troubleshooting.
Install For Coders
From PyPI
python -m pip install Orange3-CanvasAgent
python -m Orange.canvas --force-discovery
From Source
git clone https://github.com/pantagram1031/orange3-canvasagent.git
cd orange3-canvasagent
python -m pip install -e ".[dev]"
python -B -m unittest discover -v
python -m Orange.canvas --force-discovery
First Run
- Open Orange.
- In the left widget toolbox, find the Canvas Agent category.
- Drag Canvas Agent onto the canvas.
- Open the Setup tab.
- Click Check Setup.
- If needed, use Sign in to authenticate the installed
codexCLI. - Click Test Agent.
- Open the Chat tab and send a request such as:
Add a File widget and a Data Table widget, then connect them.
- Review the proposed changes on the Preview tab.
- Click Keep Changes to accept them or Revert AI Commit to undo them.
Troubleshooting
- Installer did not find Orange: Re-run the installer and use Choose Folder... to select the Orange installation directory.
- A black command window flashed: Install
v0.1.3or newer from the latest release. The wizard now hides subprocess windows during install, verification, and launch. - Installed, but Canvas Agent is not visible in Orange: Close Orange,
reopen it, and check the left widget toolbox for the Canvas Agent
category. If you installed from PyPI or source, run
python -m Orange.canvas --force-discoveryonce, then reopen Orange. - The widget category is still missing: Re-run the Windows installer if you
used
CanvasAgentSetup.exe, or verify that your PyPI/source install went into the same Python environment used by Orange. - Codex CLI missing: Install the
codexCLI first, then return to the widget and click Check Setup again. - Codex CLI signed out: Use the widget's Sign in button or run
codex loginin a terminal. - Agent request failed: Open the Diagnostics tab and copy the report before filing an issue.
Screenshot Placeholders
[Placeholder] Installer welcome and Orange detection screen[Placeholder] Installer finish screen after successful install[Placeholder] Orange toolbox showing the Canvas Agent category[Placeholder] Canvas Agent Setup tab[Placeholder] Canvas Agent Preview tab with Keep/Revert controls[Placeholder] Diagnostics tab with discovery report
What The Widget Does
- Adds a dedicated Canvas Agent widget category to Orange.
- Uses structured canvas actions instead of arbitrary model-generated Python.
- Keeps an in-memory checkpoint so each AI change can be accepted or reverted.
- Includes Setup, Chat, Preview, and Diagnostics tabs.
- Checks both Codex CLI readiness and Orange widget discovery state.
Release files for Orange3-CanvasAgent 0.1.3
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| orange3_canvasagent-0.1.3.tar.gz | 38.3 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| orange3_canvasagent-0.1.3-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 56.4 kB
Release files / orange3_canvasagent-0.1.3.tar.gz
| Download URL | orange3_canvasagent-0.1.3.tar.gz |
|---|---|
| Size | 38.3 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
3807743b3ba95d25b80aedafc74cf3d1a0f1381e82aecf4e1c2bfbb62f3ee3f1
|
|
BLAKE2b-256 checksum How to use checksums |
07949395c42492934c0b6799e9a52e4db3fd4007d906dfa258eba6dd927ad91f
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on May 12, 2026.
Transparency logRelease files / orange3_canvasagent-0.1.3-py3-none-any.whl
| Download URL | orange3_canvasagent-0.1.3-py3-none-any.whl |
|---|---|
| Size | 18.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
c7dbd73357ca77ab5fcc06b81111c69a3a6a9cd57cf4433bbe446fe9efaa2fd3
|
|
BLAKE2b-256 checksum How to use checksums |
bf8b19375979af4b95d525b5f27569e026d74c9665be99e8a92ecaf0b008c5ce
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on May 12, 2026.
Transparency log