---
title: "Prompt improver"
method: POST
path: "/api/v1/ai/prompt-improve"
tags: ["ai"]
---

# Prompt improver

`POST /api/v1/ai/prompt-improve`

## Request

- Base URL: `https://toolforte.com`
- URL: `https://toolforte.com/api/v1/ai/prompt-improve`
- Auth: one of:
  - HTTP bearer
  - none

Rewrite a rough prompt for an AI model into one that gets better results: a clear role, the task, the context the model is missing, constraints, the output format, and an example when that helps. Returns the improved prompt, a list of what changed and why, and tips for iterating. Works for ChatGPT, Claude, Gemini and open models; say which one if it matters. Costs 10 credits per call with an API key or a signed-in session, drawn from the same monthly balance as REST calls (1 credit) and renders (25). Without either, a small free demo allowance per day. Also available on the MCP server as improve_prompt.

## Request body

- object
  - `prompt` string, required: The prompt as you would type it now.
  - `goal` string: What a good answer looks like, if the prompt does not say.
  - `model` string: The model it is for, for example ChatGPT, Claude, Gemini.
  - `style` 'concise' | 'detailed', required: concise keeps the prompt short; detailed adds structure and an example.

## Response `200`

The tool's answer plus metering

- object
  - `ok` boolean
  - `data` object
    - `improvedPrompt` string, required
    - `changes` string[], required
    - `tips` string[], required

## Other responses

- `400`: Missing or invalid request fields
- `429`: Daily demo allowance or monthly credits used up
- `502`: The model did not return a usable answer
- `503`: AI temporarily unavailable

---

[API](https://skmtc.dev/toolforte/apis/toolforte-api.md) · [All operations](https://skmtc.dev/toolforte/apis/toolforte-api/llms.txt) · [OpenAPI document](https://skmtc.dev/toolforte/apis/toolforte-api/revisions/4d814d484dfc?raw)