Stacktrace AI
Stacktrace AI provides a central control plane for organizations running AI coding agents, offering visibility into every agent session, policy-based guardrails, and real-time alerts when behavior drifts. Its Fleet product aggregates findings from all agents—local or remote—into a single dashboard, while the Personal Agent Gateway verifies and controls the personal agents customers bring to a company's product. The platform includes default detections for leaked credentials and retry loops, with alerts delivered via Slack or Claude Code.
- Developer Tools
- Artificial Intelligence
- AI Agents
- Cybersecurity
- Software Only
Funding
Raised to date
$500KRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Founder details are not available yet.
Product
Problem
AI coding agents increasingly run unattended in CI, cloud environments, and on schedules, holding credentials while performing hundreds of tool calls over extended sessions. As agents multiply and run longer, misbehavior and drift go unobserved, and early mistakes compound into larger failures that no person can review or approve.
Solution
Stacktrace provides a central control plane that gives organizations visibility into every agent session, pre-set policy limits on what agents may do, and real-time alerts that flag when behavior crosses a boundary. The Fleet product aggregates findings from all agents a team runs, whether local or remote, into a single dashboard. The Personal Agent Gateway verifies and controls personal agents that customers bring to a company's product, holding risky actions for customer approval. Slack and Claude Code integrations deliver alerts as they happen, and policy rules are compiled into each environment.
Target Audience
Primary customers are engineering teams running AI coding agents in production or CI environments, and SaaS companies whose products are accessed by end users' personal AI agents.
Features
- Default detections for leaked credentials and runs stuck in retry loops, with identification of affected sessions and blast-radius analysis
- Policy engine that lets teams choose which components agents can use, with per-environment compilation
- Inventory tracking of MCP servers, plugins, and skills, including how they change over time
- Continuous vulnerability monitoring of components against known advisories
- Real-time alerts delivered to Slack or Claude Code, with agent session views in a unified dashboard
- Personal Agent Gateway capability that verifies agent identities, blocks known names from wrong domains, and holds sensitive tool calls (e.g., refunds) for customer approval
- In-browser sandbox for checking a session locally without uploading files to Stacktrace servers