Templates
Community starter packs for convilyn-author. Discover, install, and fork ready-made workflows so you can ship in minutes instead of starting from a blank server.py.
Templates live under the convilyn-template-* namespace on PyPI. The CLI verbs template list, template install, and template fork are at sdk-author/src/convilyn_sdk/cli/main.py:395-471.
Browsing the catalog
convilyn-author template list
convilyn-author template list --query analyzer
Each line is a published package name — convilyn-template-* — that you can install verbatim:
convilyn-template-doc-analyzer
convilyn-template-meeting-notes
convilyn-template-pdf-extractor
Source: cli/main.py:407-431. The --query flag is a case-insensitive substring filter on the suffix (the part after convilyn-template-). Brand-new namespaces may return:
No templates found. The convilyn-template-* namespace is brand-new;
ask the community channel for a starter or publish your own under
that prefix.
This is a feature, not a bug — the empty state tells you the namespace is real but waiting on community submissions.
Options
| Flag | Default | Notes |
|---|---|---|
--query <text> | none | Case-insensitive substring filter on the template suffix |
Installing a template
convilyn-author template install doc-analyzer
Source: cli/main.py:434-449. The argument is the suffix — the CLI prepends convilyn-template- for you, so install doc-analyzer resolves to pip install convilyn-template-doc-analyzer.
Expected output:
Installed convilyn-template-doc-analyzer==0.2.1
Source: https://github.com/CoreNovus/convilyn-template-doc-analyzer
The install is recorded in your local registry at ~/.convilyn/templates.json so later commands (e.g. template fork) can resolve the package's source repository without re-querying PyPI.
Forking + customizing
When you want a starter that's yours to evolve:
convilyn-author template fork doc-analyzer my-doc-pipeline
Source: cli/main.py:451-470. The fork command:
- Locates the source repo from
~/.convilyn/templates.json(or PyPI metadata if not installed yet). - Clones it into
./my-doc-pipeline/. - Rewrites the package name, server slug, and workflow
spec_idso nothing collides with the upstream template.
Expected output:
Forked to /path/to/my-doc-pipeline
Next:
cd my-doc-pipeline
pip install -e .
convilyn-author dev
The three follow-up commands install your fork in editable mode and start the local dev loop — from there you iterate exactly like you would on a hand-written project.
Publishing your own
Submit a template by publishing a PyPI package whose name starts with convilyn-template- and pointing the [project.urls].Source field at the public repo. The template list discovery walks the PyPI index for that namespace prefix, so a fresh release is browsable within the index refresh window (usually under an hour).
Minimum pyproject.toml:
[project]
name = "convilyn-template-doc-analyzer"
version = "0.1.0"
description = "Document analysis starter for convilyn-author"
requires-python = ">=3.11"
dependencies = ["convilyn-sdk>=1.0"]
[project.urls]
Source = "https://github.com/your-org/convilyn-template-doc-analyzer"
Inside the package, ship a server.py + workflow.py shaped the same way Tool Servers and Workflows describe — template fork does the renaming, so the source repo can keep its own slugs.
Related
- Hosted Runtime — deploy the forked template once you've made it your own
- Tool Servers — shape of
server.pyinside a template - Workflows — shape of
workflow.pyinside a template