Getting started

Your first prompt, three ways

1. In the browser

  1. Open the workspace and paste a prompt, or pick one from Templates….
  2. The Analysis tab updates as you type. Expand any finding to see why it matters and how to fix it.
  3. Choose a mode (conservative, balanced, structured) and a goal, then press Optimize.
  4. Review the changes, the diff, and the candidates table, then choose Use this or copy the result.
  5. Save keeps the prompt in this browser's local storage. Nothing is uploaded.

2. From the command line

The CLI runs the same engine offline. Until the npm packages are published, build it from the repository:

shell
git clone https://github.com/vijayrajpalanisamy88/Prompt.git && cd Prompt
corepack enable && pnpm install
pnpm --filter @promptflow/engine build && pnpm --filter @promptflow/cli build
node packages/cli/dist/cli.js analyze my-prompt.md
node packages/cli/dist/cli.js optimize my-prompt.md --mode structured

See CLI & CI gate for every command.

3. Over HTTP

curl
curl -s https://promptflowengine.com/api/v1/analyze \
  -H 'content-type: application/json' \
  -d '{"prompt":"Write something about our product.","model":"gpt-4.1"}'

See HTTP API for endpoints and limits.

Reading the results

  • Quality is 100 minus a fixed penalty per finding. See scoring.
  • Tokens are exact for OpenAI models, approximate for others, and heuristic until the tokenizer loads.
  • Cost is an estimate from reference prices. Override the price in Settings if yours differs.