Busabase × ChatGPT

Knowledge base for ChatGPT

ChatGPT can find, compare, and draft an answer in minutes. Busabase gives that answer a home your whole team can use: records with fields and sources, Docs and files beside them, and a Change Request on every edit ChatGPT makes — so you can see what changed, who made it, and undo it.

Connect ChatGPT to Busabase over MCP in a few minutes. Claude Code, Codex, and your other agents read and write the same workspace.

Last verified 13 November 2026

Connect ChatGPT How Change Requests work

The short version

What ChatGPT is great at

Searching, reading files and connected apps, comparing options, and drafting the answer inside a conversation.

What Busabase adds

A durable team knowledge base: visual tables, Docs, files, and Skills in one workspace, with grep across all of it and full change history.

How they connect

Add Busabase as a custom MCP server in ChatGPT (https://busabase.com/api/mcp), sign in with OAuth, and pick the key level: read, propose, or write.

A useful answer is not yet a durable fact

A chat can combine web results, uploaded files, and connected apps. It is still shaped for a conversation. A record your team runs on needs a stable identity, fields, evidence, an owner, and a history of who changed it.

Where it lives	Best for	What it keeps
A ChatGPT conversation	Explaining and exploring	Narrative, tied to one chat
A ChatGPT project	Keeping related work together	Useful background inside the project
A Busabase record	Coordinating downstream work	Structured fields, sources, owner, and every change as a Change Request — readable by people and any agent

Every change ChatGPT makes is a Change Request

ChatGPT can ask before an app runs a write action — that protects the tool call. Busabase records what the call actually changed. Every edit ChatGPT makes to a record, Doc, or Skill arrives as a Change Request: a message, a field-level before and after, the author and the key it came through, and a one-step rollback.

You choose how much ChatGPT can do. A read key lets it search and fetch. A proposal-only (changeRequest) key means its changes wait in the Inbox for a person. A write key lets its changes merge at once — still with the full diff and undo.

A message and a diff on every write. Each change says what it does and shows the field-level before and after.
Know which agent did it. The author, and the key or tool the change came through, stay on the commit.
Undo is built in. Records, Docs, and Skills keep their full history; merging never overwrites, and any change can be rolled back.
Review when you choose. Writes with write access merge at once. Give an agent a proposal-only key, or send a change with autoMerge: false, and it waits in the Inbox for a person to approve or send back. Read the Change Request docs.

A Busabase Change Request under review: its message, a before-and-after diff of a record's body, and Approve or Request changes controls — A Change Request in the Inbox: what changed, who proposed it and through which tool, and the before and after of the field.

Two different questions, two layers: ChatGPT's confirmation asks “may I call this tool now?”; the Change Request answers “what exactly changed in our data, and can we take it back?”

From research to a record your team reuses

When a research memo becomes a supplier decision, a launch brief, an account update, or a policy exception, pull the claim and its source into fields. The record should make sense without reopening the chat.

Gather. ChatGPT reads the sources and apps you select for the message.
Distil. It separates facts, inferences, open questions, and the recommended action.
Write. It creates or updates the record in Busabase; the edit lands as a Change Request with its message and diff.
Review where it matters. For ledgers that carry consequences, a proposal-only key holds the change for the owner to approve or send back.
Reuse. People, apps, and other agents read the current record instead of an old chat.

A small research ledger

Store what another person needs to verify, not every sentence ChatGPT wrote. This compact schema works for decisions derived from research and connected business data:

Field	Purpose
subject	The company, account, market, policy, or decision being tracked
claim	One concise statement the team may rely on
source_urls	Primary evidence and connected-record references
confidence	Confirmed, supported, uncertain, or contradicted
owner	The person accountable for the record
effective_at	When the fact starts to apply
next_check	When it should be checked again

A Busabase Base showing agent-written posts as a table with title, cover image, and body columns, plus saved views — A Base in Busabase: records an agent wrote, readable by anyone on the team. Saved views filter the same records.

ChatGPT alone, or ChatGPT with Busabase

Rows are the jobs teams bring to a ChatGPT knowledge base. The first row is ChatGPT's home ground; Busabase does not replace it.

What you are trying to do	ChatGPT with Busabase	ChatGPT alone
Research, summarise, and draft from the web and your files	ChatGPT does this; Busabase supplies the team's records and Docs as context	Yes — its home ground
Keep findings as structured records with sources and owners	Yes — Bases with real field types	Answers stay in conversations and projects
See every change, who made it, and undo it	Yes — a Change Request on every edit	Not designed for record-level history
Let teammates browse and fix the knowledge	Yes — table, gallery, kanban, calendar, and Gantt views	Shared chats and projects
Find exact text across files, Docs, and records	Yes — one grep with line and context	Search within ChatGPT
Use the same knowledge from Claude Code, Codex, and scripts	Yes — one workspace over MCP, CLI, and API	Tied to ChatGPT
Turn the knowledge into an app or a website	Yes — AirApps, including websites on your own domain	Not a ChatGPT feature

What else the knowledge base gives you

A team knowledge base

Docs, files, and Bases sit in one tree with one permission model. Members join with read, write, or manage roles; agents join with their own keys.

grep across everything

One regex search over files, Doc bodies, Base records, and agent prompts, returning the line and its context. How grep works.

Shared team Skills

Upload your team's Skills once and every connected agent reads the same version. Editing a Skill is a Change Request, so you can see who changed an instruction and why.

AirApps on live data

Ask an agent to build a small web app on your workspace data. It runs in the browser, on this machine, in a cloud sandbox, or as a website.

Websites on your own domain

The Website engine runs on Cloudflare Workers, can have backend endpoints, and takes a custom domain (Plus and above on Cloud). Agents preview and publish through the CLI, API, or MCP. Website docs.

Agents connect in minutes

MCP, a CLI, an OpenAPI REST API, webhooks, a Python SDK, and plugins for Claude Code, Codex, and DeepSeek Harness that bundle the official skills. Connect your agent.

Open source, runs anywhere

MIT-licensed. Run it on a laptop, self-host it, run it offline or air-gapped, or use Busabase Cloud. A space exports to plain files: Markdown, Base definitions, and JSON records.

Flat, predictable pricing

Free plan for small teams; paid Cloud plans are a flat price per seat with no metered overage. Self-hosting is free with no seat limit. Pricing.

Start from a template

Ready-made workspaces such as CRM and content operations, with Bases, Docs, and Skills already set up. Templates.

Connect it, then scope it

Connect. In ChatGPT on the web, add a custom MCP server with https://busabase.com/api/mcp, choose OAuth, and sign in to Busabase. Step-by-step guide.
Workspace controls on the ChatGPT side. OpenAI's docs say “Account and workspace policies apply to adding and using custom MCP servers,” and that “As an admin, you decide who can use each plugin and what its apps can do.”
Record controls on the Busabase side. The Busabase identity you sign in with decides which spaces and records ChatGPT can reach, and the key level decides whether its edits merge or wait. As OpenAI puts it, “Asking for confirmation doesn’t replace restrictions on what the app or connected account can access.”
Private servers. ChatGPT cannot reach localhost. OpenAI documents Secure MCP Tunnel “to connect a private MCP server in ChatGPT without exposing the server to the public internet.” Busabase Cloud uses its public HTTPS endpoint, so no tunnel is needed.

Coming next

These are being built now. They are not available yet, and this page will say so until they ship.

App usersIn development

People who use an app you publish sign in with an email code and see only their own records. Enforced by the platform; they are not workspace members and take no seat.

OntologyIn development

Define business objects — Company, Contact, Deal — and the links between them on top of your existing Bases, so people and agents share one vocabulary. Changes to it go through Change Requests.

Fine-grained permissionsIn development

Row- and field-level access inside a Base, agent grants scoped to a task with an expiry date, and access requests that a person approves.

ChatGPT knowledge base FAQ

Does Busabase replace ChatGPT memory or projects?

No. Those help ChatGPT keep conversational context. Busabase holds structured records, Docs, and files with sources, owners, and change history that your whole team and every agent can reuse.

Can ChatGPT write directly into Busabase?

Yes, when ChatGPT's workspace settings allow write tools and the Busabase key allows it. With write access its changes merge immediately and still carry a diff and undo. Give it a proposal-only key when you want each change to wait for a person.

Does every change need approval?

No. Review is optional. Use a proposal-only key for ledgers that carry consequences, and write access where speed matters more. Either way, every edit is a Change Request you can read and roll back.

Why not paste the whole research report into one field?

A long report is good evidence, but downstream work needs stable fields — claim, owner, effective date, confidence, next check. Attach or link the report and extract only what needs reuse.

Can ChatGPT connect to Busabase running on my own machine?

Not directly; ChatGPT cannot reach localhost. OpenAI documents Secure MCP Tunnel for private servers. Busabase Cloud works over its public HTTPS endpoint.

How is this different from ChatGPT's own shared pages?

See Busabase vs ChatGPT Space for a side-by-side comparison.

Start with one record

Connect one workspace, pick a small decision ledger, and let ChatGPT write the first sourced record. Then open its Change Request and see exactly what it did.

Connect ChatGPT Browse templates

Sources: OpenAI: Connect and test your plugin, OpenAI: plugin and app administration