NEW
Connect Claude, Cursor, or any MCP agent in 60 seconds·
Connect Claude, Cursor, or any MCP agent in 60 seconds·
Connect Claude, Cursor, or any MCP agent in 60 seconds·
Connect Claude, Cursor, or any MCP agent in 60 seconds·
Connect Claude, Cursor, or any MCP agent in 60 seconds·
See how
MCPAgents propose. You approve.

The governed context layer forAI agents.

Poggle gives your AI agents a workspace they can read — and a trust gate they can't cross. Agents connect over MCP, read your context, and submit changes as proposals. Nothing is written until a human approves it.

Your notes never train AI models
Encrypted at rest & in transit
Export everything as Markdown, anytime

Free to start · No credit card required

poggle · AI EditsConnected
Pending review3 in queue
Claudeproposednow

UpdateAPI Architecture

## Rate limits
+ 1,000 requests / min per token
+ Burst: 50 / sec, then 429
- Legacy per-IP quota note
+123Engineering
Audit trail
  • Every decision is logged, immutably.

Compatible with

  • ClaudeClaude
  • ChatGPTChatGPT
  • CursorCursor
  • CopilotCopilot
  • WindsurfWindsurf
  • GeminiGemini
  • ClineCline
  • KimiKimi
  • GrokGrok
  • DeepSeekDeepSeek
  • QwenQwen
  • ManusManus
  • HermesHermes
  • OpenCodeOpenCode
  • Kilo CodeKilo Code
  • VS CodeVS Code

Poggle exposes a standard MCP server — if your agent speaks MCP, it connects. No SDK, no lock-in.

The loop

One governed loop, end to end.

Connect, read, propose, approve. Every agent follows the same path, and a human holds the gate at the end of it.

Step 01

Connect over MCP

Any MCP-capable agent connects to your workspace with a scoped token. One protocol, no bespoke integrations to maintain.

Agent
scoped token
Workspace
The trust gate

Give agents context.
Keep control of the truth.

Letting an AI agent write straight to your knowledge base is how small mistakes become permanent ones. Poggle puts a human between the agent and your source of truth — without slowing the agent down.

Read freely, write never. Agents have rich read access but cannot mutate anything directly.

Approve in context. Review each proposal as a diff, then approve, edit, or reject.

Reverse anything. Full version history and one-click rollback on every object.

See the full flow
CClaude · access
Scoped
  • ReadEngineering · Support · GTM
  • ProposeEngineering
  • Write directlyNever
  • DeleteNever
Every action lands on an append-only audit log.
Why Poggle

Governance built in, not bolted on.

Proposals, not writes

Agents submit reviewable diffs. You see exactly what would change, in context, before it happens — no silent edits to your source of truth.

A real trust gate

Approval is required, not optional. Scope what each agent can read and propose, and keep a human in the loop on every change.

Every action audited

Connect, read, propose, approve — it's all on an append-only log with full version history and one-click rollback.

Pricing

Start free, govern at scale.

Connect your first agent for free. Upgrade when you need unlimited agents and a full audit trail.

Connect your first agent and try the trust gate.

  • 1 agent connection
  • 3 boxes
  • Proposal review
  • 7-day history
Pro
Popular
$12/month

or $9/mo billed annually

For teams running agents on real context.

  • Unlimited agents & boxes
  • Scoped MCP access
  • Full version history
  • Append-only audit log
Team
$39/month

or $29/mo billed annually

Contact sales

Shared, governed context for the whole org.

  • Everything in Pro
  • Shared workspaces
  • SSO / SAML
  • Priority support

All paid plans include a 14-day free trial. No credit card required.

Full pricing details →

Put a human in the loop.

Give your agents the context they need — and the trust gate they can't cross. Free to start.