CodeIntegrity raises $4.8M seed to put runtime guardrails on agentic AI
What's the deal? San Francisco-based startup CodeIntegrityDealroom has a profile for this one. Try Dealroom → has raised a $4.8M seed round to build security guardrails for agentic AI applications. Cybersecurity-focused Syn VenturesDealroom has a profile for this one. Try Dealroom → led the round, with participation from existing pre-seed investors AntlerDealroom has a profile for this one. Try Dealroom → and Boost VC, bringing total funding to $5.25M.
The five-person company, founded in May 2024, offers a "runtime control layer" that sits between AI agents and enterprise systems. It forces unpredictable AI models to follow strict, deterministic rules — limiting which systems and data an agent can touch.
Why now? Enterprises are racing to deploy AI agents that can autonomously perform tasks, but these tools are inherently unpredictable. Unlike traditional software, where input X always produces output Y, AI agents are driven by language models and vulnerable to "prompt injection" attacks — where malicious text tricks an agent into exposing sensitive data.
CodeIntegrity made headlines last year by compromising Notion's AI features in under four hours, earning a mention in The Economist. "Every company was aiming to launch agents into deployment, and they didn't know how to do it safely," said co-founder and chief technology officer Abi Raghuram.
Current workarounds — human oversight or using a second large language model as a judge — don't scale well and aren't foolproof.
What could go wrong? CodeIntegrity faces a crowded and fast-forming competitive field. Seattle's CertivDealroom has a profile for this one. Try Dealroom → emerged from stealth in March, and California-based Raven and Manifold Security are also tackling the problem. About six companies are piloting CodeIntegrity's product, but it hasn't finalised its pricing model yet.
The signal: Syn Ventures is a cybersecurity-specialist fund, and its decision to lead a seed round for a company barely a year old underscores how quickly agentic AI security is being treated as a standalone investment thesis rather than a subset of broader AI or cyber bets. With competitors like Certiv, Raven, and Manifold Security all emerging in recent months, capital is flowing into a category that essentially didn't exist before enterprises began piloting autonomous agents at scale.
Read more: GeekWire
