These scripts illustrate the stable Python API that slackwright
exports from slackwright.__init__. Anything imported here is part of
the public surface and will not break across 0.x patch releases.
For agents (LangChain, dspy, custom orchestrators, MCP wrappers) that already run inside a Python process, embedding directly skips the subprocess + JSON-parse round-trip and gives you typed access to the result objects.
To run any example:
# from the repo root, after a one-time `./slackwright install`
.venv/bin/python examples/01_fetch_in_process.pyThe examples assume you've already run ./slackwright login --workspace https://your-workspace.slack.com interactively at least once, so the
state directory has a valid auth bundle.
01_fetch_in_process.py— minimal end-to-end fetch: open a client, build a SearchPlan, stream messages into an ArchiveWriter. Mirrors whatslackwright fetchdoes internally.02_resolve_only.py— resolve a name / email / Slack id to a fully populated UserRecord. Useful if you only need identity lookups.03_explain_no_network.py— build a SearchPlan and the chunk schedule entirely from cached data, no Playwright spawn. The same fast-path the CLI's--explainuses.04_render_report.py— point the HTML renderer at an existing output directory and producereport.html.