Migrate your meetings from Read AI

The short version

You don't have to read this page — your AI assistant can. Copy this and give it to Claude, ChatGPT or whichever assistant has your connectors:

Migrate my meetings from Read AI into Scalign. Use Scalign's help tool to read the guide "Migrate your meetings from Read AI", then follow the prompt under "Move them over" step by step, stopping where it tells you to wait for my OK.

What this does

Switching from Read AI? Your past meetings come with you. Each one lands in Scalign as if Scalign had recorded it — summary, actions, objections, stakeholders, Reasons to Buy, deal health — so account history is there from day one. Nothing to install: your AI assistant reads the meetings from Read AI and hands them to Scalign, with both connected to it.

Why the assistant and not a button: Read AI knows who was in the meeting, but not always which voice is whose, and Scalign needs an email on every voice to place it on a person. The assistant matches them, looks a person up when the name isn't enough, and asks you when it can't — before the meeting goes in.

Coming from Fathom, Fellow, Fireflies, Modjo or Attio instead? Those have a one-click import: see Import meetings from another tool.

Before you start

• Read AI is connected to your assistant. Read AI has its own MCP server — in Read AI, open Integrations and follow their MCP setup for Claude or ChatGPT (EU workspaces use a separate server address). It needs a Read AI plan that includes API and MCP access.
• Scalign is connected to your assistant with your own account. Follow Connect Claude, Cursor and other AI tools (MCP) or Connect ChatGPT if it isn't. The meetings are imported as yours.
• You can see the meetings in Read AI. The assistant only sees what your Read AI account does. A colleague's meetings are migrated from their account, the same way.

Move them over

Open a new chat in your assistant and paste the prompt below. It lists your Read AI meetings for the period you pick, does one for your approval, then does the rest.

Migrate my meetings from Read AI into Scalign. Work through these steps in order and stop whenever a step says to.

1. Check that you have both the Read AI and the Scalign connectors. If either is missing, stop and tell me how to add it.

2. Ask me which period to migrate (for example "the last 3 months") and whether to include only meetings with people outside my company — that is the default; internal meetings are usually noise in Scalign.

3. In Read AI, call list_meetings for that period and show me the list: title, date, participants. Mark the ones you plan to skip (internal, under 5 minutes, no transcript) and wait for my OK.

4. Dry run on the first meeting. Fetch it with get_meeting_by_id. Build the Scalign transcript from its transcript: one turn per speaker utterance in time order, with the speaker's name, their email, the text, and start and end in seconds from the start of the meeting (Read AI gives timestamps in milliseconds from the Unix epoch: subtract the meeting start and divide by 1000; if an utterance has no end, use the next one's start). For the emails:
   - Match each speaker label to a participant's email by name.
   - If a speaker is only "Speaker 1" or the participant has no email, first try to find the person elsewhere: the meeting's invite list, or Scalign's list_stakeholders for the company (the domain of the other participants). Say what you found.
   - If you still can't, ask me who the speaker is. If I don't know either, drop that speaker's turns and tell me how much of the meeting that was.
   - Use one spelling of each name per email, and one email per name.
   Use the real meeting link as meeting_url if Read AI has one. Show me the title, the start time, the speaker-to-email mapping and the first few turns, and wait for my OK.

5. After my OK, call import_call_recording with it. Then do the rest of the list the same way without asking, except for speakers you cannot place — collect those and ask me about them in one go, then import. Importing the same meeting twice is harmless: Scalign returns the existing recording.

6. End with a summary: meetings imported, meetings skipped and why, and any speakers you dropped. Analysis takes a few minutes per meeting; I'll see them under Meetings in Scalign.

Throughout: never invent an email, and never split a meeting across several imports unless it is over 60,000 words.

What ends up in Scalign

One recording per meeting under Meetings, owned by you and shared the way your recordings normally are. The source shows as import. Analysis runs within a few minutes of each import; companies, stakeholders and deal health follow from it like for any recorded meeting.

If Scalign also recorded one of the meetings and the assistant passed the meeting link, the transcript attaches to that recording instead of making a second one. Importing the same meeting twice returns the existing recording.

If something is off

• A speaker is on the wrong person — the email on their turns was wrong. Fix it on the meeting in Scalign: see Wrong or missing speaker names: how Finalize works.
• "Too few words for analysis" — the meeting was too short to say anything about. Skip it.
• "At most 60,000 words" — a very long workshop. Ask the assistant to split it at a natural break and import the parts as separate meetings.
• "A recording of this meeting exists that … doesn't have access to" — a colleague recorded it privately. Ask them to share it, or tell the assistant to import without the meeting link.
• A meeting is missing from the list — Read AI's list is paged, 10 at a time, and the search covers about the last year. Ask the assistant for the next page, or name the meeting.

Still stuck? Email support@scalign.ai.