Build with Reshaper

Three ways in, one name

Whichever you pick, the name is always reshaper-ai.

# use it inside Claude or Codex
npx reshaper-ai@latest install-mcp

# use it from your terminal
npm install -g reshaper-ai

# start a project built on the API
npm create reshaper-ai@latest

Most people want the first one. The second is for scripts and your terminal. The third scaffolds a small app wired to the API, for building a product on Reshaper.

Quick start

Create a key on your API keys page, then send it as a bearer token. Every endpoint below uses the same header.

curl https://reshaper.ai/api/v1/reshape \
  -H "Authorization: Bearer rai_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "text": "we shipped dark mode today",
    "inputType": "Raw Thoughts",
    "outputType": "LinkedIn Post"
  }'

{
  "outputType": "LinkedIn Post",
  "notes": ["Dark mode shipped today. Here is what it taught us..."],
  "creditsSpent": 1,
  "creditsRemaining": 47
}

Text output always comes back in notes, an array. Some formats return more than one piece from a single call, so treat it as a list.

Inside Claude and Codex

The MCP server gives an AI assistant 47 Reshaper tools: write a post, generate every format, render a carousel, turn a reel into writing, optimize a YouTube video, read and save to your library, schedule what it writes, and teach it a brand voice from your own documents. It works with your voice and your credits, and it can read files off your machine.

npx reshaper-ai@latest login rai_live_...
npx reshaper-ai@latest install-mcp

claude mcp add --transport http reshaper https://reshaper.ai/api/mcp \
  --header "Authorization: Bearer rai_live_..."

{
  "mcpServers": {
    "reshaper": {
      "command": "npx",
      "args": ["-y", "reshaper-ai", "mcp"],
      "env": { "RESHAPER_API_KEY": "rai_live_..." }
    }
  }
}

Once it is connected you can just ask: "turn my last YouTube video into a LinkedIn post and a carousel, in my brand voice."

From your terminal

npm install -g reshaper-ai
reshaper-ai login rai_live_...

reshaper-ai reshape "we shipped dark mode today" --out "LinkedIn Post"
reshaper-ai all @notes.md --save
reshaper-ai carousel "5 lessons from 2 years of building" --template black
reshaper-ai youtube https://youtube.com/watch?v=... --thumbnail
reshaper-ai reel ~/Downloads/talk.mp4 --out "LinkedIn Post"

Every command that takes text also accepts - for stdin and @file.txt for a file, so it pipes like any other Unix tool. Run reshaper-ai help for the full list.

Sending files

Anywhere the API wants a photo, video, PDF or document, you can send it three ways, whichever is easiest: an https URL, a data: URI, or raw base64. Generated images always come back as URLs.

You: make a carousel from ~/Downloads/talk.mp4

Claude calls create_carousel with
  audio: "/Users/you/Downloads/talk.mp4"

The bridge strips the video, sends the audio,
and you get the finished slide URLs back.

# a local video: audio is extracted first
reshaper-ai reel ~/Downloads/talk.mp4 --out "LinkedIn Post"

# no ffmpeg? hand it a Drive link instead
reshaper-ai reel https://drive.google.com/file/d/.../view

# teach a workspace its voice from a brand deck
reshaper-ai clients docs <id> ./brand-guide.pdf

# teach Reshaper YOUR voice from your own writing
reshaper-ai voice add ./post-1.txt
reshaper-ai voice add ./post-2.txt
reshaper-ai voice add ./post-3.txt
reshaper-ai voice analyze

Endpoints

Base URL https://reshaper.ai. Costs are in credits.

Call GET /api/v1/meta for the live list of valid input types, output types, carousel templates, cover style ids and prices. It needs no key, so you can load it at build time.

Scheduling

Write a post and queue it, in one go. Scheduling runs through Buffer and costs no credits — you are only placing text you already generated.

# 1. find a channel
curl https://reshaper.ai/api/v1/buffer \
  -H "Authorization: Bearer rai_live_..."

# 2. queue the post
curl https://reshaper.ai/api/v1/scheduled-posts \
  -H "Authorization: Bearer rai_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Dark mode shipped today...",
    "channelId": "5f2b...",
    "dueAt": "2026-09-20T09:00:00Z"
  }'

# or from the CLI
reshaper-ai schedule channels
reshaper-ai schedule add @post.txt --channel 5f2b... --at 2026-09-20T09:00:00Z

In Claude this becomes one sentence: "turn this into a LinkedIn post and schedule it for Tuesday at 9am." It checks your channels, resolves the time, writes the post and queues it.

In your own product

const res = await fetch("https://reshaper.ai/api/v1/reshape", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.RESHAPER_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    text: transcript,
    inputType: "YouTube Video",
    outputType: "LinkedIn Post",
    clientId: acmeWorkspaceId, // optional: use that client's voice
  }),
});

const { notes, creditsRemaining } = await res.json();

Pass a clientId from GET /api/v1/clients to write in that workspace's brand voice. Leave it out and your own personal voice is used instead.

Credits and errors