Local automation runner with built-in LLM steps – YAML pipelines
YAML pipelines with built-in LLM steps running locally instead of GitHub Actions.

YAML-configurable event pipelines with AI transforms when Segment and Zapier dominate.
Backend engineers and data teams
Segment · Zapier · Customer.io
So I built my own, re-usable solution that handles the above. You build a pipeline, throw events at it (SDK or HTTP), and it runs on every event. In the middle of the pipeline you can enrich the payload with an HTTP call, branch/filter, reshape fields, or run an AI step when you need one. Then it fans out to your destinations. The whole pipeline is also representable as YAML, so it's reviewable and diffable rather than a flowchart you click together. Agents can have a token and fire data at a pipeline for further processing.
The sources and destination catalogue is still small tbh. but I'm adding more as I go.
It's live at ingestlayer.com. I'd appreciate any feedback, especially what would make you go "I'd use this if it also did X."
YAML pipelines with built-in LLM steps running locally instead of GitHub Actions.
FFmpeg-shaped pipeline orchestration for LLMs with built-in JSON validation and repair.
Visual IDE for data transforms challenges dbt's YAML-heavy workflow directly.
LLM-as-judge metrics beat guessing chunk sizes, but Ragas and LangSmith already exist.
Yet another CI/CD, but Woodpecker, Drone, and Jenkins already solved this space.
Cursor for data transformations using Ibis to standardize AI-generated SQL pipelines.