Ag2Trust

BETA

AI agents you can trust in production.

Bring your LLM keys. We provide the secure runtime with hard cost limits, tool governance, automatic PII redaction, and PII-safe audit logs — whether you're running agents internally or embedding them in your product.

ag2trust — agent config
$ curl -X POST https://api.ag2trust.com/v1/agents \
  -H "Authorization: Bearer $AG2_API_KEY" \
  -d {
    "name": "support-agent",
    "model": "gpt-4o",
    "team": "customer-success",
    "cost_limit_daily": 25.00,
    "tools": ["search_kb", "create_ticket"],
    "pii_redaction": true
  }
# ✓ Agent created with PII protection, $25/day budget, 2 tools

Powered by

AWS KMS Docker Presidio OpenAI Anthropic Bedrock

Platform Capabilities

You bring the LLM. We provide the secure, compliant agent runtime — enterprise-grade governance at a price built for startups and growing teams.

Hard Cost Limits

Daily token budgets, per-run limits, and fail-closed enforcement. Never overspend—agents stop when limits are reached.

Automatic PII Protection

All content—ingress and egress—is redacted using Presidio-powered detection. Audit logs are PII-safe by default.

Comprehensive Audit Trails

Every agent interaction, user action, and system event is logged. PII-safe by default with configurable retention policies.

Agent Teams

Organize agents into teams with communication hierarchies. Agents collaborate, transfer tasks, and request revisions—all with guardrails.

Knowledge Base + RAG

Connect Google Drive, SharePoint, or upload files directly. Automatic chunking, embeddings, and semantic search so agents answer with your organization's context.

Custom Tools

Define your own tools via API and let agents call them. Allowlist per agent, enforce rate limits, and audit every invocation automatically.

MCP Tool Integration

Connect any MCP-compatible tool server. Allowlist specific tools, assign to teams, and grant access per agent type.

Webhooks & Events

Trigger agents from GitHub, GitLab, Linear, Stripe, or custom webhooks. HMAC verification and automatic task routing.

Per-Customer KMS Encryption

Dedicated AWS KMS keys per customer with AES-256-GCM envelope encryption. Your credentials are isolated and deletable on demand.

Enterprise-Grade Security, Built In

You shouldn't need a Fortune 500 budget to ship agents safely. Your API keys stay yours, your LLM provider, your choice — we provide the secure agent runtime with sandboxed tool execution and PII-safe audit logs.

  • Per-customer AWS KMS keys (AES-256-GCM)
  • Non-root container execution
  • Network isolation with dual-network architecture
  • Multi-layer rate limiting
  • PII-safe audit logs with configurable retention
  • Toxicity detection and content moderation
  • SSRF protection on all external URLs
  • reCAPTCHA Enterprise for dashboard access
  • Role-based access control (RBAC)
  • Webhook HMAC signature verification

Built with Ag2Trust

See what teams are shipping on the platform.

BocaDesk

bocadesk.com

AI-powered help desk for internal teams and customer support. Employees get instant answers grounded in company knowledge bases — with PII redaction and audit trails handled by Ag2Trust.

Simple, Transparent Pricing

Pay only for what you use. No hidden fees, no surprises. LLM token costs are billed directly by your provider—not by us.

Hourly Monthly

Free

$0/month

Get started

  • 1 running agent
  • 2 agent types
  • 5 knowledge docs
  • 5 custom tools
  • No teams
  • No integrations
Start Free

Starter

$4.99/mo base

+ $4.99/agent/mo

  • Up to 20 running agents
  • 10 agent types
  • 40 knowledge docs
  • 20 custom tools
  • No teams
  • No integrations
Get Starter
POPULAR

Pro

$9.99/mo base

+ $9.99/agent/mo

  • Up to 50 running agents
  • Unlimited agent types
  • 100 knowledge docs
  • 100 custom tools
  • Teams
  • Integrations & Webhooks
Get Pro

Enterprise

COMING SOON

Custom

Tailored to your needs

  • Unlimited everything
  • Custom retention
  • SSO/SAML
  • Dedicated support
Contact Us

Prices do not include LLM token costs, which are billed directly by your provider (OpenAI, Anthropic, AWS Bedrock, etc.).

Request Early Access

We're onboarding a small number of teams — whether you're running internal agents or building AI-powered products for your customers. Request access and we'll get your workspace set up.

Quality over scale: We're intentionally limiting early access to ensure every deployment meets our security and reliability standards. Safety and trust come first.

Prefer email? Write to hello@ag2trust.com.