Workflow · saas · workflow
Val Town's journey fast-following AI code assistants into Townie: from Copilot to Claude 3.5 Sonnet
Val Town users demanded a state-of-the-art LLM code generation experience from day one, but early implementations — a ChatGPT cosplay autocomplete and a function-calling Townie — were slow, inaccurate, and had poor feedback loops that made iterating on generated code difficult.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · User describes app in Townie chat
A user describes an app in natural language and Townie produces a deployed app in about 30 seconds.
Tools used
TownieCodeiumClaude 3.5 SonnetClaude ArtifactsChatGPTGPT-3.5
Outcome
Val Town shipped a Claude 3.5 Sonnet-powered Townie in August 2024 that can generate a fullstack app with frontend, backend, and database in minutes, fully deployed, and contributed a novel automatic error detection feature.
What failed first
The ChatGPT-powered autocomplete occasionally forgot its role and gave inaccurate completions, and the function-calling Townie hallucinated non-existent functions while its generic interface was ill-suited for iterating on code.
Results
Time savedminutes to generate just a couple hundred lines of code
Running sinceAugust 2024
Grounding & classification
Source type: technical build writeup
21 fields verified against source quotes.
code generationconversational aicode diff prbuilder submittedfailure mode describednamed customerproduction runtime claimedtools describedworkflow describedsoftwarecycle time reductiontechnical build writeupai draft human approval