Skip to content

All Integrations

ChatGPT + Steady

Read and write Steady data directly from ChatGPT.


ChatGPT integration for Steady

A lot of work already starts in ChatGPT. Steady’s MCP server brings your check-ins, goals, and team activity into the same place, so the conversation you start in ChatGPT can end with an update posted to Steady, no app-switching required.

The Goal Story workflow is a nice example. Ask ChatGPT what’s changed since your last update, have it pull your previous Goal Story updates from Steady, and let it suggest how to frame what’s new. By the time you’re done thinking through it out loud, the update is ready to post — and ChatGPT can post it for you.

What ChatGPT can do with your Steady data

Surface live data inside any thread. The Daily Digest, Smart Check-in responses, Goal Stories with their full update history, Echoes, team activity, insights, and absence schedules are all available to ChatGPT. Ask it to compare two weeks of check-ins, flag what’s different about this sprint’s blockers, or summarize where each active goal stands.

Send work back to Steady from the chat. ChatGPT can file new check-ins, start Goal Stories, and add updates to existing ones. Narrate what shipped and what’s next, then have it post the check-in. Or read the last few updates on a goal, talk through the trajectory, and write the next one — same thread, no context switch.

Stitch Steady into your wider connector setup. ChatGPT and Codex both run multiple connectors in parallel, so Steady sits next to Gmail, your calendar, GitHub, Notion, or whatever else you’ve wired up. Match blockers from check-ins against threads in your inbox. Pair a customer-call transcript with last week’s Goal Story update to draft a follow-up message. From Codex, cross-reference what your team reported in Steady against the diff you’re reviewing.

Installation

The Steady MCP server URL is https://app.steady.space/mcp.

ChatGPT: Add Steady as a custom connector pointing at the server URL and complete the OAuth sign-in. ChatGPT will then treat Steady as a tool it can call against in any conversation.

Codex: Add Steady to ~/.codex/config.toml under [mcp_servers.steady] with url = "https://app.steady.space/mcp", then run codex mcp login steady to authenticate over OAuth. Full instructions live in the AI Assistants doc.

Security

Each connection uses OAuth 2.1 and is scoped to the individual user. ChatGPT and Codex can read and write the same data you have access to in Steady — nothing more. Connections are per-person, so a teammate granting access doesn’t expand anyone else’s permissions.

About ChatGPT

ChatGPT is OpenAI's AI assistant, used by millions for writing, research, planning, and analysis. It's available as ChatGPT (web, desktop, and mobile) and Codex (terminal and IDE), and supports the Model Context Protocol so it can read and write data from your team's tools — including Steady — inside any conversation.

Frequently asked questions

How does ChatGPT work with Steady?

Once you’ve added Steady as a custom connector in ChatGPT, you can ask ChatGPT to fetch your Daily Digest, summarize your team’s check-ins, draft a goal update from prior context, or post a new check-in — all from inside the same chat thread you’re already using for everything else.

What does setup look like?

In ChatGPT, add a custom connector pointing to Steady’s MCP server at https://app.steady.space/mcp. ChatGPT walks you through OAuth sign-in to grant access. The full walkthrough is in the AI Assistants doc.

Can I keep Steady alongside other connectors in the same conversation?

Yes. ChatGPT can talk to multiple MCP servers in a single conversation, so you can ask it to combine your Steady context with calendar events, Slack threads, or anything else you’ve connected.

How are permissions handled?

Steady’s MCP connection is personally scoped via OAuth 2.1 — ChatGPT inherits the exact permissions you have inside Steady. There’s no shared service account or expanded access, and each teammate connects their own ChatGPT to their own Steady credentials.

Does this work with Codex for engineering teams?

Yes. Codex supports MCP and connects to Steady the same way as ChatGPT. Add Steady to your Codex config and authenticate over OAuth, then engineers can pull check-ins, post goal updates, and combine Steady data with whatever they’re working on in the terminal or IDE.

More in AI Tools

Ready to go Steady?

See for yourself how Steady cuts coordination overhead.