# CompFly AI

> The Trust Control Plane for the Agentic Economy. CompFly provides the infrastructure to discover, validate, secure, and govern autonomous AI agents from code to production — closing the "Trust Gap" between what agents can do and what enterprises can verify, constrain, and prove.

Primary site: https://compfly.ai

## Reading this site as an agent

Every page is available as clean Markdown. Fetch it any of these ways:
- Recommended: append `.md` to any page path, dash-joining nested segments into one flat file, e.g. https://compfly.ai/why-compfly.md and https://compfly.ai/solutions-observability.md (the home page is https://compfly.ai/index.md).
- Append `?view=agent` to any page URL, e.g. https://compfly.ai/why-compfly?view=agent
- Fetch the flat twin directly under /md, e.g. https://compfly.ai/md/solutions-observability.md

Human visitors get the normal rendered page (or `?view=human`, the default).

## Core Concepts

- **Trust Control Plane**: A control point between agents and the models, tools, other agents, and data they call — it intercepts, decides, and enforces policy in real time.
- **Agent Simulation (Eval)**: Automated red-teaming and reliability testing for agents in CI/CD; findings become enforced runtime guardrails.
- **Adaptive Runtime Governance (Protect)**: Inline, per-request enforcement with cryptographic agent identity (Ed25519 DID), model/tool sanctioning, and a Model Gateway.
- **Observability (See & Prove)**: Continuous discovery of agents, behavioral drift detection, and a tamper-evident, identity-attributed audit ledger.
- **Shift-Left Security**: Security checks (PII, prompt injection, behavioral drift) embedded into the CI/CD pipeline.

## Key Pages

- [Why CompFly?](https://compfly.ai/why-compfly): The Trust Gap (Observation, Control, Auditability), the path to Controlled Autonomy, and the business case.
- **Solutions** (one control plane across the agent lifecycle, a continuous loop where Agent Simulation, Adaptive Runtime Governance, and Observability harden each other):
  - [Observability](https://compfly.ai/solutions/observability): Discover, observe, and prove every agent, shadow-AI discovery, risk tiering, live sessions, behavioral drift, and a tamper-evident audit ledger.
  - [Agent Simulation](https://compfly.ai/solutions/agent-simulation): Red-team and evaluate agents before and after they ship, adversarial and safety simulation, depth modes, defense-ceiling coverage, in CI/CD.
  - [Runtime Governance](https://compfly.ai/solutions/runtime-governance): One adaptive inline control point on every model, tool, and agent-to-agent call, decided before execution, with cryptographic agent identity.
    - [Sample Assurance Report](https://compfly.ai/solutions/runtime-governance/sample-report): Interactive AI Agent Assurance Report for the Portfolio Management Agent, control categories, per-control pass rates with CIs, residual risk register, and framework mapping.
  - [Industries + Roles](https://compfly.ai/solutions/industries): Scenarios by sector (banking, retail, manufacturing, automotive) and role (security, platform, risk & compliance).
- [Manifesto](https://compfly.ai/manifesto): Our vision for the agentic era and the end of blind trust.
- [About](https://compfly.ai/about): Who we are and why CompFly exists.
- [Careers](https://compfly.ai/careers): Open roles building the immune system for agentic AI.
- [Book a Demo](https://compfly.ai/book-demo): See CompFly govern your agents end to end.

## Resources

- [Partners](https://compfly.ai/resources/partners): CompFly's partner ecosystem — model providers, infrastructure, data platforms, and startup programs (Anthropic, NVIDIA, MongoDB, Kenexai, Google for Startups). Partnerships: partnership@compfly.ai
- [Blog](https://compfly.ai/resources/blog): Technical deep dives into agentic security. Posts:
  - [The State of Agentic Security Tools in 2026](https://compfly.ai/resources/blog/state-of-agentic-security-2026)
  - [Threat and Reliability Simulations for Agents](https://compfly.ai/resources/blog/threat-and-reliability-simulations)
  - [Why LLM Safeguards Fall Short in the Agentic Era](https://compfly.ai/resources/blog/safeguards-fall-short)
  - [Introducing Crosswind: Security Evaluation for AI Agents](https://compfly.ai/resources/blog/introducing-crosswind)
  - [CompFly AI Joins the Coalition for Secure AI (CoSAI)](https://compfly.ai/resources/blog/compfly-joins-cosai)
  - [Controlled Agentic Autonomy: A Defense-in-Depth Imperative](https://compfly.ai/resources/blog/agentic-controlled-autonomy)
- [Newsroom](https://compfly.ai/resources/media/newsroom): CompFly AI company news, announcements, and press.
- [Podcasts](https://compfly.ai/resources/media/podcasts): Conversations on controlled autonomy featuring the CompFly team.

## Knowledge Base (definitions for agents)

- [What is the Trust Gap in AI Agents?](https://compfly.ai/learn/trust-gap-ai-agents)
- [What is Indirect Prompt Injection?](https://compfly.ai/learn/indirect-prompt-injection)
- [Agentic Governance](https://compfly.ai/learn/agentic-governance)

## Legal

- [Security Policy](https://compfly.ai/security)
- [Privacy Policy](https://compfly.ai/privacy)
- [Terms of Service](https://compfly.ai/terms)

## Contact

- Book Demo: https://compfly.ai/book-demo
- Email: assist@compfly.ai