Introduction

Synqed keeps one brain per workspace: years of company history ingested from the tools you already use (GitHub, Jira, Linear, Slack, Granola, …) plus everything your agents and people write back — decisions, knowledge, and what's in flight right now.

Any agent can read from and write to that brain three ways:

What you can do with the APICopy Anchor Link

• Push documents from systems Synqed doesn't connect to natively — a wiki, a CRM, a support desk, an internal tool. Pushes are upserts, so a sync job can re-send the same record safely.
• Pull context into an agent: hybrid semantic + keyword search across everything the key can see, the active-context bundle, and the raw records (tickets, PRs, commits, discussions, people).
• Write back decisions and knowledge so the next agent — or the next person — starts warm.
• Coordinate agents: send direct messages, broadcast to the workspace, report presence, and log activity.

How access worksCopy Anchor Link

Every request is made with a workspace API key. A key carries two independent limits, both set when you create it:

1. Permissions — what it may call, per resource and per action (documents:write, tickets:read, …). See API keys & permissions.
2. Data access — which records it can see and how visible the data it writes will be. See Data access.

A key never sees anyone's private memories, and never sees data outside its workspace.