Gateway that lets a vendor's AI agents reach a customer's private apps, databases and APIs through one MCP endpoint, with each customer setting what is allowed, what needs approval and what is blocked, and a log of every request.
Stacktrace
New· 4Reads coding agents' session logs on the machine to flag leaked credentials and runaway runs, enforces component policies, and can steer or stop an agent; a gateway does the same for customers' agents. Team tier free, 200 session hours a month.
- Cybersecurity
- DevOps
- Access Control
- Artificial Intelligence - AI
TrustList summary
Not yet independently verified. No maker launch date, so it is dated from the Y Combinator launch listing. The 21 detections, the claim that transcripts stay on the machine and the founders' backgrounds are the company's own; we have seen no independent test, and neither page states where it is based. We will update this when it can be confirmed, and remove this note.
Coding agents increasingly run unattended on long tasks, and permission to act is not the same as a reason to. Stacktrace, a Fall 2026 Y Combinator company, supervises those sessions while they run, with a second product aimed at the agents customers bring to a business.
Watching sessions where they happen
Fleet reads each agent's own session log on the machine, and the company says transcripts stay there. It shows every tool call, the tokens a session used against its usual, and an inventory of the MCP servers, plugins and skills each agent loads, with continuous checks for known advisories. Twenty-one default detections cover things such as leaked credentials, runaway sessions and runs stopped by a quota. Each rule can be set to off, detect or block, and findings can go to Slack, a webhook or a SIEM. Active steering, which corrects or stops a drifting run, is part of the Enterprise tier.
A gateway for customers' own agents
Agent Gateway sits in front of a company's MCP server, verifies each agent's identity, asks the customer for an explicit yes on risky actions and keeps a masked record of every decision. The server keeps its own scopes.
Billing by agent time
The Team tier is free per organization with unlimited users and 200 session hours a month, counting only time spent running model turns or tools. Enterprise and Gateway are custom and billed annually.
Founders from infrastructure
Vinod led Apache Mesos, Micheal ran infrastructure and AI platforms at Airtable, and Prasanth spent 15 years at Amazon, according to the launch post.
Key facts
Maker: Stacktrace AI, Inc. Founders: Vinod, Micheal and Prasanth (surnames not given) Launched: 6 November 2026 on Y Combinator's launch board Pricing: Team free (200 session hours a month); Enterprise and Gateway custom, billed annually Platforms: works with coding agents including Claude Code and Cursor; MCP servers Website: stacktrace.ai
Categories & features
- Cybersecurity
- DevOps
- Access Control
- Artificial Intelligence - AI
Trust Score
An earned signal from verification, reviews, awards, transparency and engagement - the vendor can't buy it.
Updated 11/7/2026
Request a demo or quote from Stacktrace
Reviews
Write the first review of Stacktrace
Used it? Your experience helps other buyers decide.
Write a reviewQuestions & answers
No questions yet. Be the first to ask about Stacktrace.
Stacktrace alternatives
Similar listings buyers compare against this one.
Open-source database and workspace where AI agents and people keep records, docs, skills and apps; free tier, paid plans from $10 per seat a month.
SDK that finds mistakes production AI agents keep repeating and injects scoped, previewed corrections only in the situations you choose, then measures the effect against held-out sessions, for any agent framework.
An AI coworker inside Slack that completes tasks proactively using the team's own tools
Turns a webpage's actions into a structured format that AI agents can use, for developers
An open-source, code-first language for durable, auditable AI workflows used by engineering teams
Platform for building AI agents with a no-code builder for domain experts and a hosted runtime for developers
Git-based storage for the skills, tools and context that AI agents rely on
More on TrustList
Everything here links back to the same verified catalogue. Pick your next stop.
- Write a reviewHelp other buyers decide
- CompaniesAgencies, consultancies and IT service providers, ranked by verified reviews.
- ArticlesGuides, comparisons and research from the editorial desk and the community.
- AwardsAnnual recognition decided by verified reviews and an independent jury.
- LaunchesNew products and releases, voted up by the community every day.
- AI ModelsBenchmark scores and community ratings for every major model.
- RequestsBuyers describe what they need; vendors respond directly.
- PeopleReviewers, authors and makers with public profiles.
- ComparePut up to four listings side by side before you shortlist.