Claude
Add the remote integration, authenticate, enable the tools you trust, and Claude uses the same private context service — under the scopes you grant it.
Agent compatibility
One context service. Your choice of agent.
Your assistant already speaks the language. BEFORE plugs into ChatGPT and Claude through the open MCP standard, so they retrieve the minimum relevant context on demand — without copying your history into every conversation.
ChatGPT
Connect the BEFORE app, grant scopes, then ChatGPT retrieves context when its tools are enabled for the conversation and surface.
Pair
Set up your band in the BEFORE app once.
Connect
Add the BEFORE connector in ChatGPT or Claude — the same way you add any integration.
Choose scopes
Grant exactly the context domains you want. Change or revoke anytime.
What “available” means
Connected agents can search and fetch the minimum relevant context through explicit tools. Platform behavior, plans, installation and review rules still apply. We will not market silent injection into every chat because neither platform guarantees that behavior.
What the connector can provide
- Ideas, decisions, commitments and open loops from intentional captures.
- Project and person context with timestamps and source links.
- Full transcripts only when separately allowed.
- Health summaries and series only for hardware-accepted measurements, with source, unit, freshness and confidence.
- Candidate actions that another agent connector can prepare or perform after approval.
Synthetic example
Your AI asks the connector: “What did I decide about the launch?”
Decision · captured Thu 17:42
“Founding batch ships first — general availability only after hardware validation.”
Voice noteScope: extracted context
Actions remain accountable
In the first release, BEFORE supplies the intent and context. Sending a message, changing a calendar, buying something or submitting a document remains an explicit action inside the user's agent and its authorized service.
Your AI is ready. Are you?
Order your band