Beta version: no payments, Pro is free on request. Some features may change. Report a problem
echoflow

echoflow connectors

Use echoflow from your AI assistant: once connected, you ask it to process a video, wait for the content, read and edit it, and ask questions about all your past videos, without opening the site. Connectors are included in every plan, free trial too. Choose where you want to use it:

What it is, in short

A connector is how an AI assistant uses an external app on your behalf. echoflow's is an MCP server (Model Context Protocol), the open standard AI assistants use to connect to services: you don't need to know how it works, you just connect it once.

You connect it with your echoflow account, with no keys to create and no configuration to write: choose the workspace, authorize, and the assistant works with your videos and your minutes. The same connector works for every assistant below.

Claude (claude.ai and apps)

  1. In Claude open Settings → Connectors and press “Add custom connector”.
  2. Name it Echoflow and paste this address:
https://api.echoflow.it/api/v1/mcp
  1. Press Connect: echoflow opens. Sign in, choose the workspace and authorize.
  2. In a chat, turn on Echoflow from the tools menu and ask for what you need.

Once Echoflow is in Claude's connector directory, you'll just search for it and press Connect.

ChatGPT

  1. In ChatGPT open plugins, press Add and choose a custom MCP server.
  2. Name it Echoflow and paste this address:
https://api.echoflow.it/api/v1/mcp
  1. Press Connect: echoflow opens. Sign in, choose the workspace and authorize.
  2. In a chat, turn on Echoflow and ask for what you need.

Once Echoflow is in ChatGPT's app directory, you'll just search for it and press Connect.

Claude Code

In your terminal, add the server with this command:

claude mcp add --transport http Echoflow https://api.echoflow.it/api/v1/mcp
  1. Start Claude Code and type /mcp.
  2. Choose Echoflow and complete the sign-in in your browser.

From Claude Code you can also upload video and audio files from your computer.

OpenAI Codex

The command line, the editor extension and the desktop app share the same configuration. In your terminal, add the server with this command:

codex mcp add Echoflow --url https://api.echoflow.it/api/v1/mcp
  1. Your browser opens: sign in to echoflow, choose the workspace and authorize. If it doesn't open, run codex mcp login Echoflow.
  2. Start Codex and ask for what you need, for example “Use Echoflow to list my latest videos”.

To upload a file from your computer, Codex needs network access: approve the command when it asks.

Other AI assistants

It works with any MCP-compatible assistant that supports OAuth sign-in: use the same address and sign in with your account.

For an assistant without OAuth, or for your own integration, you need an API key instead, included in the Pro plan: instructions are on the Developers page.

What you can ask

Write as you would to a person: the assistant works out what to do. The examples work in Claude, ChatGPT, Claude Code and Codex, and each one uses the features of your plan.

Processing a video

  • “Process this YouTube video and write the LinkedIn and Instagram posts.”
  • “Upload interview.mp4 from my Downloads folder and wait for the transcript.”
  • “How many minutes would a 45-minute video cost, and how many do I have left?”
  • “Process this podcast with subtitles and vertical clips for reels.”
  • “The last video didn't go through: retry from where it stopped.”
  • “Add a note to yesterday's video: publish on Friday.”

Reading, editing and downloading

  • “Give me the full transcript of the latest video.”
  • “In the blog post, fix Ranuci to Ranucci and save.”
  • “Also generate the newsletter for this video.”
  • “What are the best moments and quotes from Monday's video?”
  • “Translate the summary and the posts into English and Spanish.”
  • “Give me a link to download all the video's files in a single zip.”

Archive and AI Chat (from the Pro plan)

  • “What did I say about pricing in the last three webinars? Give me the minutes.”
  • “Search all my videos for where I talk about onboarding.”
  • “Pick up yesterday's AI Chat conversation and sum up the open points.”

Writing and planning

  • “Rewrite the summary of the latest video in a more direct tone and save it.”
  • “Give me some video ideas about remote work and write the script for the best one.”
  • “Plan a two-week editorial calendar from the latest episode.”
  • “In the calendar, move the LinkedIn post from Tuesday to Thursday.”

Your style

  • “Set my Brand Voice: direct tone, plain words, an audience of accountants.”
  • “Remember that I always sign my posts with my name at the end.”
  • “Change the LinkedIn post prompt so posts always end with a question.”
  • “In the Brand Kit, upload logo.png and use our brand blue as the main color.”

Sharing and account

  • “Create a public link to share this video's summary.”
  • “How many minutes have I used this month and how many videos have I processed?”
  • “Switch to client Rossi's workspace and show me their latest videos.”

Pro plan

  • “Build a 30-day campaign from the latest webinar.”
  • “Prepare September's monthly report for the client, with the commentary.”

What the connector does

The same things you do in the web app: start a video from a link or a file, with the subtitle style and the clip format, follow its processing, read, edit, translate and download the content, AI Chat and archive search, Rewrite, Ideas, editorial calendar, Brand Voice, AI Memory, Brand Kit, custom prompts and share links. On the Pro plan also the AI Marketing Assistant campaigns and the white-label monthly report.

It works on one workspace at a time: ask the assistant to switch to another workspace you belong to, for example a client's, and carry on with the same requests there.

Payments, workspace members, API keys, company sign-in (SSO), two-factor authentication and deleting the account or workspace stay in the web app only.

Before an action that deletes or cancels something, the assistant asks you to confirm.

Plans and minutes

The connector is included in every plan, free trial too, at no extra cost. Every request follows your plan's rules as in the web app: processing a video uses minutes, and each feature stays in the plan that includes it (see pricing). After the trial, without a plan, you can browse the archive but not process new videos.

Your data

The assistant receives only the content you ask it for: for example the transcript of a video when you want a summary. It is data you send to a service you chose, under your own agreement with its provider (Anthropic for Claude and Claude Code, OpenAI for ChatGPT and Codex); echoflow does not sell or share your content with anyone else.

When connecting you choose the workspace and what to grant: reading, creating and processing, browsing the archive. You can revoke the connection at any time from Settings → Connected apps: it stops working immediately.

The rest is in the privacy notice. Questions: privacy@echoflow.it.

Support

If something doesn't work, write to us from the Help center, saying which assistant you use (claude.ai, the Claude app, Claude Code, ChatGPT or Codex).