Bidirectional data redaction
Detect and mask CPF, CNPJ, SSNs, Chinese IDs, payment cards, emails, JWTs, API keys, and database URLs in prompts and responses.
AegisMCP is the zero-trust security gateway between your LLMs, AI agents, and internal APIs. Redact PII in both directions, inject ephemeral credentials in memory, and enforce granular access on every MCP tool call.
$490/month after the trial · Credit card requiredKeep sensitive data and credentials behind a policy-controlled boundary.
Detect and mask CPF, CNPJ, SSNs, Chinese IDs, payment cards, emails, JWTs, API keys, and database URLs in prompts and responses.
Agents request an action, never a password. A permitted credential is decrypted only for the upstream call and is not returned to the model.
Apply YAML rules to agent roles, tools, and actions. Deny destructive operations or financial actions by default.
Every allow, block, and redaction is recorded in an organization-scoped SHA-256 hash chain.
The gateway checks the identity, plan, tool, and arguments before anything reaches an internal API.
Sensitive values are tokenized and allowed credentials are injected only inside the gateway.
The decision and redacted result are written to an audit chain for review and export.
Every self-serve plan includes the full product for seven days. A card is collected at checkout; billing starts on day eight unless canceled.
No. AegisMCP is a paid production service. Self-serve plans start with a seven-day full trial and require a payment card at checkout.
Stripe collects your card at signup. Your selected subscription is charged automatically after the seven-day trial unless you cancel in the customer portal.
Growth includes Docker and Helm self-hosting. Enterprise adds air-gapped deployment options.
AegisMCP provides controls and audit evidence that can support SOC 2, LGPD, GDPR, HIPAA, and PIPL programs. It does not itself certify your organization.
$490/month after the trial · Credit card required
Start 7-day trial ↗