Multi-stage AI pipeline

Turn any input into finished work.

Foundry runs your content through a sequence of AI stages — research, narrative, editorial — and produces a finished piece. You control every step.

Get startedSign in
How it works
01
Give it a topic.
Paste any input — a transcript, article, idea, or URL. Select a pipeline and a lens to frame the output.
02
Watch it run.
Each stage executes in sequence. Research → Narrative → Editorial. You see every output as it's produced and can intervene at any step.
03
Get finished work.
Review, edit, and export. Foundry tracks every run so you can refine your pipelines over time.
Built for professionals

Your keys. Your models. Your pipeline.

Foundry runs on your Anthropic and Perplexity API keys. No markup, no data retention, no black box. You configure each stage independently — model, provider, prompt — and the pipeline executes exactly as you’ve defined it.

BYOK
Bring your own Anthropic and Perplexity API keys. Direct billing. No intermediary.
Custom pipelines
Build pipelines with any stage sequence. Override prompts per stage.
Lenses
Apply a lens to any run — a context layer that shapes every stage's output.

Ready to build your pipeline?

Set up takes under five minutes.

Get started