manifest-generator

Automating project manifest generation involves scanning Machina workflows to aggregate credentials, connectors, datasets, and agents into a draft document. This process extracts dependencies statically to streamline setup and reduce manual errors.

The best way to extract dependencies for a manifest is through static extraction, which automatically identifies and aggregates credentials, connectors, datasets, and agents directly from your workflows without manual input.

Yes, you need MongoDB and google-genai to generate project manifests. MongoDB stores the aggregated data while google-genai provides optional LLM enrichment to enhance the manifest with additional information during generation.

Yes, you can enrich a draft manifest using an LLM. The generation process supports optional LLM enrichment via google-genai, allowing you to enhance the statically extracted dependencies with additional contextual information.

To generate a manifest draft for a specific template, run a POST request to the workflow executor endpoint. This triggers the aggregation of dependencies and outputs a draft manifest document ready for review and commit.

Manifest generation outputs a draft document for review. It aggregates dependencies and optionally enriches them with an LLM, but the final manifest requires manual review and commit before deployment.