Choose your workshop path

Choose how you want to participate. We recommend inspecting the saved conversations in your own Agent Canvas. You can follow the recorded decisions and tool results without waiting for a new run or paying for model calls.

Option	What you do	When to choose it
1. Read the HTML	Open the saved transcripts in your browser.	Quick access with no Canvas setup.
2. Inspect in Agent Canvas (recommended)	Import the lecture histories, filter the conversation list, and expand tool calls and results.	Follow along with the instructor and investigate the traces yourself.
3. Run the full exercises	Start fresh conversations and perform the tasks with your own model access.	You have time to configure the tools and compare new runs.

Download the lecture examples ZIP and extract it. The selection contains 13 lecture conversations, three worker histories and two original animation outputs.

Read the HTML

Open the HTML examples online, or open harness-lecture-conversations/package/index.html from the extracted ZIP. If local files are restricted, run python3 serve_transcripts.py inside the extracted folder and open the printed URL. No model key is needed.

Inspect the conversations in your own Canvas

These directions are for a local npm installation of Agent Canvas 1.29.0, using Agent Server 1.56.0. The import was tested on macOS with a fresh Canvas store. Windows, Linux, Docker and remote-backend imports have not been verified; use the HTML option if this local setup does not match yours.

1. Install the tested Canvas version, or check your existing version:

   npm install -g @openhands/agent-canvas@1.29.0
   agent-canvas --version

   Follow the Canvas installation prerequisites if Node.js or uv is missing.

2. Stop Agent Canvas with Ctrl-C in its terminal. Keep the backend stopped during import.

3. Open a terminal in the extracted harness-lecture-conversations folder. Preview the import, check the printed destination, then install:

   python3 import_into_canvas.py
   python3 import_into_canvas.py --install

   No model key or Canvas session credential is needed. The importer checks the package hashes, refuses conflicting conversation IDs, and adds 16 histories: 13 lecture conversations and three team-worker histories. It preserves their 526 recorded events. Repeating the same import verifies the installed files without duplicating them.

4. Restart Canvas:

   agent-canvas

   Open the URL printed in its terminal. On first launch, choose your usage-data preference and Skip for now on model setup. The “LLM isn’t set up” banner only affects new runs; saved histories still open. Beside Conversations, choose View → Filter by tags → lecture=odsc2026. Click More if the list is collapsed. Open a conversation and expand its tool calls and observations.

The npm launcher normally stores conversations in ~/.openhands/agent-canvas/dev_conversations. If you configured another local store, pass its absolute path with --store /absolute/path/dev_conversations. If your backend uses a custom port, pass its loopback URL with --backend-url http://127.0.0.1:PORT so the stopped-backend check uses the right port. The importer previews by default and changes files only with --install.

The imported records are for inspection. To try a task yourself, start a fresh conversation in the exercise's workspace. Historical commands refer to the recorded environment, and imported archive agents have no model credentials or executable tools configured.

The earlier stale-memory example uses historical custom tool types. The importer supplies documented generic compatibility wrappers for those records; messages, tool-call arguments, results, IDs, and timestamps are retained. The original package remains unchanged, and odsc-lecture-import.json beside the store records the import and compatibility mappings.

Run the full exercises

If you have time, follow the exercise prompts and model/tool setup. Start fresh conversations in the specified workspaces. New results may differ from the recorded examples.

Recover or remove the import

Keep the original ZIP and extracted package. If the importer reports a collision or changed history, keep your existing data; it does not overwrite it. The receipt lists the exact imported IDs and event hashes. Stop Canvas before removing imported history folders, and remove only those listed in the receipt. Leave other conversations, settings, credentials and workspaces alone.