Control exactly which tools see what
Every memory in Reflect has an allowed_vendors field that determines which AI tools can read it. Set granular access per memory, use wildcards for broad sharing, or lock context to a single agent key — all manageable from the dashboard or API.
Precise access for every memory
Visibility controls let you decide exactly which AI tools and agents can read each piece of context — from a single vendor lock to a team-wide wildcard.
"allowed_vendors":["cursor", "claude"]
Per-memory vendor scoping
Each memory carries its own
allowed_vendors list. When an AI tool requests context, Reflect checks the requesting vendor against that list before returning anything. No match, no data.*
cursor
claude
Wildcard or explicit access
Use
* to share a memory with every connected tool, or list specific vendor slugs for tight control. Defaults are configurable per workspace so new memories inherit your team's policy automatically.Memory access, your way
Share selectively across tools
Give Cursor access to your coding standards while keeping HR policies visible only to ChatGPT. Each memory's vendor list is independent, so you share exactly what each tool needs.
Keep sensitive context private
Credentials, internal architecture notes, and compliance data can be locked to a single tool or agent key. Visibility controls enforce boundaries that API keys alone cannot.
Audit who accessed what
Every memory read is logged with the requesting vendor and key identity. Pair visibility controls with the audit trail to get a full picture of which tools consumed which context.
Fine-grained control over your AI memory
Start scoping memory access per vendor in minutes. Free tier includes full visibility controls.