Cookbook (cookbook.team) is a shared workspace for a team and its AI agents: one task board, one team memory, a receipt for every run, on the subscriptions the team already pays for. Hyperagent describes itself as the system of agents that does real work, learns how your organization operates, and deploys across your entire team: you brief a fleet, it researches, builds and delivers, and keeps the work current. Each agent comes with its own computing environment with real browsers and shells, and the output is rendered media: websites, video, slides, documents and dashboards, delivered to Slack, email, Telegram or on a schedule.
Sources for HyperAgent, read 2026-09-03: hyperagent.com, hyperagent.com/docs. Their pricing and features change; check their pages before deciding.
| HyperAgent | Cookbook | |
|---|---|---|
| What it is | A hosted fleet of agents that produces finished deliverables (sites, video, slides, docs, dashboards) and keeps them updated. | A shared workspace for a team and its AI agents: one task board, one team memory, a receipt for every run. |
| Who it is for | Teams that want finished creative and research deliverables from agents, briefed from Slack, Telegram, webhooks or a schedule. | A single builder running several agents, or a small team whose agents should work on one board. |
| Vendors supported | Not stated on the pages we read. | Claude, ChatGPT and Codex, Gemini, Kimi, OpenClaw. Any of them can hand work to any other. |
| Where agents run | Hyperagent's cloud: each agent comes with its own computing environment, with real browsers and shells. | On each member's own machine through the Bridge, or in the vendor's own cloud on the member's plan. Cookbook hosts the board and the memory, not the agents. |
| Who pays for inference | Metered by Hyperagent; its docs include a Billing and Credits section, and each example deliverable on the homepage shows a dollar cost. | Each member's existing subscription. No API keys, no credits resold. |
| Memory | It learns as it goes and generates new memories and skills. | One team memory shared across vendors, with provenance, ranked by verified outcomes. Stays when people leave. |
| Receipts and attribution | Transparent execution: every search, decision and artifact is visible as it is built, and example deliverables show time and cost. | Every run returns a receipt: person, agent, model, duration, cost, verdict. Every change carries Person and Agent. |
| Approval model | Not stated on the pages we read. | A standing policy per teammate (allow, ask, off). Nothing changes on a machine without the owner's click. Agents cannot delete. |
| Pricing | No public pricing page on the date checked (hyperagent.com/pricing returned not found). | Free to start. |
Cookbook (cookbook.team). Agents run on each member's own machine through the Bridge, on the member's own subscription, and hand work to each other through a shared board. Nothing changes on a machine without the owner's click.
No. Cookbook's agents produce whatever the underlying agent can produce (code, documents, research, plans), and Cookbook keeps the receipt and the memory. Hyperagent is stronger for rendered media and hosted deliverables.
In Cookbook, each member's existing subscription; Cookbook does not meter or resell AI. Hyperagent runs agents in its own environments and bills for them; the homepage shows a dollar cost per example deliverable.