granola — Codex plugin: meeting context and setup
granola connects Granola meeting history to ChatGPT, letting the assistant find decisions, requirements, and customer feedback while you work.
Open on GitHubWhat granola does
granola connects meeting history to ChatGPT through MCP. The assistant can search conversations by topic, person, company, or time period, then use the relevant material and point to the meetings behind its answer.
It can bring customer feedback and stakeholder decisions into PRDs or proposals, combine several conversations with research for reports, and recover project requirements from planning discussions while you code. Meeting context can sit alongside documents and other sources.
If the relevant conversation is not in Granola’s history, the assistant cannot cite it. The plugin is listed for Codex and ChatGPT.
What it helps with
- find customer feedback across meetings
- bring meeting decisions into a PRD
- recover project requirements from planning calls
- combine meeting context with report research
- trace when a decision was made
Example prompts for the agent
- Search Granola for meetings about this project and list the technical decisions with the conversations they came from.
- Draft a PRD from meetings with this company, citing the specific conversations behind each major requirement.
- Gather customer feedback about the new feature from past meetings and group the findings by topic.
Questions about granola
How do I install granola in Codex?
Connect granola in Codex as an MCP plugin so the assistant can access Granola meeting history.
What does granola do?
It gives ChatGPT access to Granola meetings. The assistant searches by topic, person, company, or timeframe and brings the relevant conversations into its response.
Which platforms does granola support?
granola is listed for Codex and ChatGPT.
Source text
granola · MIT
Granola MCP connects your meeting history to ChatGPT so your assistant can pull real context from past conversations while you work.
Use MCP when you're already working on something bigger:
- Writing PRDs or proposals: Pull customer feedback, stakeholder input, and decisions from relevant meetings directly into your draft
- Building presentations or reports: Combine insights from multiple conversations with your other research and data
- Coding with context: Give your AI the project background, requirements, and technical decisions discussed in planning meetings
- Strategic analysis: Merge meeting insights with documents, data, and other sources for comprehensive analysis
How it works:
Your AI assistant can search your Granola meetings by topic, person, company, or timeframe - then cite specific conversations in its responses. No more switching tabs to find "what the customer said" or "when we decided that."
The difference:
Instead of copy-pasting meeting notes into prompts, your AI pulls exactly what it needs from your meeting history automatically. You stay focused on the work, not the research or context.
© Granola · MIT · openai/plugins