| What it does | Secure AI agents, with evidence
| Catch your AI agents' mistakes, block the dangerous ones
|
|---|
| Stage | Not given
| Startup
|
|---|
| Founded | Not given
| 2026
|
|---|
| Based in | GR
| Not given
|
|---|
| Pricing model | Freemium
| Not given
|
|---|
| Pricing | Community: €0 (open source, 1x monthly testing volume, weekly monitoring, 3 seats, 30-day data retention). Team: €349/mo (20x monthly testing volume, 10 seats, 90-day data retention, webhooks, priority support). Enterprise: Custom pricing (contact for details).
| Free: $0/month with 5,000 conversations/month, Langfuse or LangSmith integration, problems and Slack alerts, rules that block costly actions, no credit card required. Starter: $49/month for 25,000 conversations/month plus email support. Growth: $149/month for 100,000 conversations/month plus priorit
|
|---|
| Key features | - Automated adversarial testing with multi-turn and agentic attack chains
- Black-box testing without SDK installation or agent modification
- Four-tier Humanbound Firewall with LLM judge
- Continuous posture score from 0 to 100
- Framework mapping to EU AI Act, NIST AI RMF, OWASP LLM Top 10, OWASP Agentic AI Top 10
- Air-gapped testing with self-hosted models like Ollama
- Export compliance evidence in HTML, PDF, JSON, SARIF, CEF formats
- Webhook integration with SIEM and ticketing systems
- Apache-2.0 open source CLI, testing engine, and SDK
- Baseline campaign completion in hours
| - Monitor AI agent conversations without code by connecting Langfuse or LangSmith
- Judge conversations on six criteria: accuracy, goal alignment, decision quality, completeness, helpfulness, and factuality
- Create code-based rules that block risky actions before they run
- Group failures into problems with suggested rules and failure counts
- Send Slack alerts for serious mistakes and reply checks immediately
- Add to agents with one line of Python code for OpenAI Agents SDK, LangChain and LangGraph
- Mask sensitive data like email addresses, phone numbers and card numbers before sending to Kalytera
- Rules continue running offline if Kalytera goes down
- Correct judge verdicts by marking conversations as passed or failed
|
|---|
| What makes it different | - Continuous monitoring instead of point-in-time reports
- Adaptive multi-turn attack chains that evolve over time
- Same open-source code as platform with nothing held back or limited
- No account required for local open-source testing
- Framework mapping and compliance evidence for regulatory requirements
| - Rules are code-based, not AI, so they cannot be overridden or argued with
- No code setup using Langfuse or LangSmith connections for immediate visibility
- Rules continue protecting agents even when Kalytera is offline
- Catches exact mistakes from real AI agent incidents like database deletes, account takeovers and crypto theft
- Customer data stays on your servers; Kalytera only reads conversations, never writes or stores sensitive data
|
|---|
| Community votes | 0 | 0 |
|---|
| DR (Domain Rating by Ahrefs) | 14 ▲ +1 (up) since last check | 1 ● no change since last check |
|---|
| Trust Flow (Majestic) | 8 | 0 |
|---|
| Citation Flow (Majestic) | 30 | 14 |
|---|
| Referring domains (Majestic) | 71 | 28 |
|---|
Humanbound