Kontext
Munich startup, runtime security for AI agents. $4M seed in 2026 led by 42CAP.
Europe's most interesting startup stories. Every week.
No spam. Unsubscribe anytime.
Podcast Episode
🎙️ Deep Dive
Metycle, 50skills, Kontext — €131.7M for recycled metals, $6M for HR AI and $4M for agent security — Daily News September 24, 2026
Listen to the episodeAbout Kontext
The Munich startup sits between AI agents and the tools they use, running as a local daemon that checks every tool call (including from coding agents such as Claude Code and Codex) against policies defined at multiple levels (organisation, group, user, agent, repository, branch). It can observe, block, or require approval before execution and records every decision for audit. The product started with coding and customer-support agents but is shifting toward supporting agents customers run themselves, with adoption concentrated mostly in financial services (mid-sized and large banks and fintechs). In September 2026, Kontext closed a $4 million seed round led by 42CAP, with participation from a16z CSX and HTGF; the valuation was not disclosed.
The Story
Founded in Munich in early 2025 by Jens Ernstberger (PhD in system security and applied cryptography, prior venture-capital experience at a16z Crypto) and Michel Osswald (electrical engineering background, moved into cybersecurity), both with backgrounds in secure computing, applied cryptography, and AI systems.
How Kontext works
Business Model
Tiered B2B SaaS (free Starter, Pro, Scale) plus enterprise contracts, aimed mainly at development and security teams in regulated companies.
Revenue Model
SaaS subscriptions: free plan for developers, $149/month for teams (Pro), $499/month for higher volume and retention (Scale), custom enterprise pricing.
Products & Services
Core Use Cases
- Pre-execution control of AI-agent actions
- Enforcement of contextual policies
- Audit trails for coding agents and cloud workloads
Market & Clients
Key Customers
Security and software-development teams; enterprises deploying AI agents, with early adoption concentrated among mid-to-large banks and fintechs.
Geographic Presence
How Kontext competes
Competitive Advantages
- Decision-making before execution
- Combined context of identity, task, and resource
- Observe mode before enforcement
How Kontext grows
Growth Strategy
Expand the engineering team; support more agents and environments; convert free developers into team and enterprise plans.
Distribution Model
Free plan for developers; SaaS subscriptions for teams ($149/month Pro, $499/month Scale); custom enterprise contracts.
Moat (Defensibility)
Multi-level, task-aware policy engine; runtime action and decision data collected in production.
Regulatory Context
EU AI Act and governance obligations; GDPR and data minimisation in logs; audit and security requirements for regulated sectors such as financial services.
Key Risks
- False positives that interrupt work
- False negatives that allow harmful actions
- Performance overhead and fragile integrations
- Competition from cloud and identity platforms
Strategic Insights
- Early customer concentration in financial services suggests AI-agent controls find their first paying demand where mistakes carry an immediate regulatory or financial cost.
- Shifting from a coding-agent focus to broader support for customer-run agents is a typical demand-driven product pivot rather than the original thesis.
Funding & Investors
Total Funding
Funding Rounds
- 2026 - : $4M
Lead: 42CAP
Key Investors
Founding Team
Founders
Key Executives
- Jens Ernstberger - CEO
Key Metrics
Lessons from Kontext
- Public, tiered pricing (free, team, enterprise) can be a credibility lever for a still-young security product, making perceived value tangible.
- In security products, starting in observe mode before enforcement can reduce adoption friction, letting policies be validated against real traffic before blocking actions.
Similar Startups
Devo
Cloud log analytics and security software born in Spain as Logtrust, now with a declared HQ in Boston. Valued at $2B in its 2022 Series F.
Read AnalysisDarktrace
Cambridge AI cyber security firm: Series E at $1.65bn in 2018, London listing in 2021, acquired by Thoma Bravo (about $5.3bn) in 2024.
Read AnalysisNeuralTrust
Barcelona startup building a platform to discover, monitor, control and protect enterprise AI agents; €17.2M seed led by Alstin Capital.
Read AnalysisSources
- Kontext raises $4M to stop AI agents from going rogue at work — Kontext
- Kontext raises $4M to stop AI agents from going rogue at work — HTGF
- Kontext raises $4M for runtime security platform for AI agents — Tech.eu
- Kontext raises $4M to put guardrails on AI agents for banks and fintechs — TechFundingNews