How to give Claude your team's meeting context
Three ways to get your team's meetings in front of Claude: paste, Projects, or a connector. What each does well, and how to set up the connector in minutes.
To give Claude your team's meeting context, you can paste transcripts into a chat, upload them to a Claude Project, or connect Claude to a source that holds your meetings through an MCP connector. The first two work for one person and a handful of calls. The third is the one that keeps working when a whole team is on calls every day.
Claude only knows what is in front of it. Without help, it has not heard the call your colleague took yesterday, and it cannot tell you what the customer said last month. Here are the three approaches, in the order most teams try them.
Option 1: Paste the transcript
Copy the transcript from your notetaker, paste it into a chat, and ask your question.
This is fine for a one-off: summarize this call, draft the follow-up email, list the action items. It costs nothing and needs no setup.
It breaks down quickly. You have to find and copy the transcript every time. Long transcripts eat up the conversation. And the next day, in a new chat, Claude has forgotten all of it.
Option 2: Upload transcripts to a Claude Project
A Claude Project lets you attach files and instructions that apply to every chat inside it. Put the transcripts for one customer or one initiative in a project, and Claude can draw on them in each conversation.
This is a good step up if one person works on one account. To get the most from it:
- Create one project per customer, deal, or initiative.
- Upload the transcripts as files, with the date and attendees in the file name.
- Add a short instruction, for example: "Always say which call a statement comes from."
- Re-upload after every new call.
The catch is step 4. A project is a snapshot that someone has to keep current. It also belongs to whoever set it up, so your teammates' calls are only there if they upload them too. And it covers meetings only, not the email thread or the calendar invite that explain them.
Option 3: Connect Claude to a shared source over MCP
MCP is the open standard that lets an agent like Claude call outside tools. A connector gives Claude a live way to search your meetings instead of a pile of files you uploaded.
This is what removes the copy-paste. The meetings are captured in one place, and Claude looks things up when you ask. If that place is shared, Claude can answer from what the whole team heard, not only what you remembered to upload.
What to look for in a connector:
- It reads and writes. Claude should be able to add what it learns, not only look things up.
- It respects permissions. Claude should see only what you can already see.
- It shows sources. An answer like "they asked for a Q4 start" needs a date and a quote you can check.
How to do it with Nysa
Nysa is shared memory for teams that work with AI agents. A notetaker joins the calls you choose as a visible participant and transcribes them, then turns meetings (and Gmail and Google Calendar) into facts with sources. Claude reads those facts through an MCP connector. Setup takes under 5 minutes.
To connect Claude:
- In Claude, go to Customize, then Connectors, and search for Nysa.
- Select Connect and sign in with the Google account you use for Nysa.
- Start a new conversation. Claude picks up the Nysa tools automatically.
If you can't find Nysa in the list, choose Add custom connector and enter https://mcp.usenysa.com/mcp.
If you use Claude Code in the terminal, run:
claude mcp add --transport http nysa https://mcp.usenysa.com/mcp
The full steps are on the Connect Claude page. Claude sees only what your Nysa account can already see, each call is audited, and changes are made only when you ask Claude to make them.
Questions to try first
Once connected, test it with questions that need real team context:
- "What did we agree with this customer on our last call?"
- "Who on the team has spoken to this company, and when?"
- "What are the open questions from this week's meetings?"
Check that each answer points back to a meeting or message. If it does not, see Saved isn't searchable for how to test what your agent actually has.
Which option should you pick?
A Claude Project can hold a few transcripts for one person. As soon as more than one person takes calls, or you are tired of re-uploading, connect Nysa: its notetaker captures the calls you choose, setup takes under 5 minutes, and Claude searches what the whole team heard. The same brain can also serve your other agents; see how to give every agent on your team the same company context.
FAQ
Does Claude remember my meetings between chats?
Not on its own. Each conversation starts with only what you paste in, what is in the project, and what a connected tool returns. A connector solves this because Claude searches your meetings fresh each time.
Can Claude see my teammates' private information through the connector?
With Nysa, Claude sees only what your Nysa account can already see. Raw email stays private to its owner, and the team gets the facts drawn from it.
Do I need to be technical to set this up?
No. In Claude you search for Nysa under Connectors, select Connect, and sign in with Google. The terminal command is only for people who use Claude Code.