For coding agents: the Studio guide
The exact texts Bicycle Studio serves to coding agents through the studio_guide MCP tool and GET /api/studio/v1/guide, synced from the Studio repository.
Working with the person
How to work with the person: questions first, confirmed numbers, person-only steps.
Start here: building on Bicycle Studio
What to build, the order of work, the rules.
Connecting to the Studio MCP
URLs, OAuth vs token, scopes, and why some tools may be missing.
Data apps
Compose vs hand-build, bda.manifest.json, bda.fn, upload and publish.
Functions: a guide for coding agents
Kinds, function.json, tests and fixtures, the code sandbox, calling a function.
Agents
Agent functions: the spec, connections, limits, the queue.
Workflows
The create, try and publish loop, steps, sends, triggers, calling by ref.
Runs: invocations and traces
One inv_ record per run: status, events, trace, debugging.
Detect and Explain: a guide for coding agents
What changed, and why: Detect and Explain on an app's model.
Limits and person-only steps
Every step only a person may take, and the limits.
Porting a data story or notebook into a workflow
Porting a notebook into a workflow: what maps to which step.
Example: a KPI function and a hand-built app that shows it
A KPI code function and a hand-built app that shows it.
Example: an Ask AI function that summarises a small table
An llm function that summarises a table.
Example: a classify function that sorts support tickets
A classify function that sorts tickets.
Example: an agent function, and the code function it will call
An agent function with one function tool.
Example: weekly bookings digest by email
A weekly workflow whose email a person approves.
Example: train a model, then predict with it
Train, then predict, with workflow blobs.
Example: a backtest mapped over origins
A backtest mapped over partitions.
Example: hand a publish to a person
A token builds an app version; a person publishes it.