Credal
4.2
Credal governs AI agents in the enterprise: an MCP platform for controlling tool calls, an agent builder, and an agent registry for tracking what exists. It enforces existing document permissions at query time, so a user cannot retrieve what they could not already open. It offers SSO and SOC 2 Type II. Pricing is custom.
- Commercial use OK
- Enterprise contract
- SOC 2
- Pricing
- Contact sales
- Free tier
- No
- Commercial use
- Permitted
- Japanese UI
- Not confirmed
- Enterprise contract
- Available
- Compliance
- SOC 2
What it is good for
- Organisations with agents proliferating faster than oversight
- Enforcing document permissions through the AI layer
- Controlling which tools an agent may call
- Security teams that need an inventory of deployed agents
Strengths and trade-offs
Strengths
- Permission inheritance at query time solves the hardest RAG governance problem
- Agent registry addresses shadow AI directly
- MCP tool-call governance is ahead of where most of the market is
Trade-offs
- No published pricing
- Only relevant once you have agents to govern
- Adds a layer between agents and data that must itself be operated
Editors' review
The problem arrives about six months after the first agent
The failure teams hit is a RAG system that cheerfully surfaces the salary spreadsheet because the index did not carry permissions. Enforcing the source system's permissions at query time is the correct fix and is harder than it sounds.
Nobody buys this before they have the problem, which means it usually arrives after an incident.
Verdict
Buy it when agents are multiplying. Design for it earlier if you can.
Compiled from official documentation and public sources, with our own editorial judgement. We do not publish figures we cannot source.