TORONTO, ON — October 16, 2026

Securetron today announced the Community Edition of AI-FW, its self-hosted AI firewall and governance gateway, now free to use with 100 million tokens per month. The release gives engineering and security teams a production-grade way to inspect, govern, and audit the traffic between their AI agents and the large language models they call — without a license key, an account, or a per-seat fee.

AI agents have become the fastest-growing attack surface in enterprise software. A single agent can hold provider API keys, touch customer data, and call external tools, yet most deployments send prompts straight to a model endpoint with no inspection, no identity check, and no audit trail. AI-FW sits between agents and models to close that gap: every prompt and response is scanned in real time for jailbreaks and prompt injection, PII is masked, toxicity is filtered, and data-exfiltration attempts are blocked — inbound and outbound.

The gateway is fail-closed by design and never retains content. All observability is built from metadata: a transaction log of what happened, not the text of what was said. That design also satisfies the hard line that stops many regulated teams from adopting AI controls — the firewall never stores the prompts or responses it inspects.

"The Community Edition is the full inspection and governance pipeline, free," said Sarah Reid of Securetron. "Teams get jailbreak detection, PII masking, semantic policy scoring, model routing, agent identity, and a metadata-only audit log for 100 million tokens a month. It is enough to secure a real deployment, not a demo."

Beyond guardrails, AI-FW brings the operational controls enterprises expect from network firewalls to AI traffic. Semantic intent analysis scores requests against natural-language policies using embeddings, a chat classifier, or a fully offline engine. Intelligent routing sends each request to the right model and backend, with per-model keys, failover, and weighted distribution across providers. Agents authenticate with API keys, JWT, mTLS, or Kerberos; roles are enforced with RBAC; and SSO connects through OIDC with SCIM provisioning.

The release also targets the quiet cost problem in AI adoption. An exact and semantic completion cache re-serves repeated prompts without another upstream call — Securetron reports an identical second request returning in 0.0075 seconds versus 0.64 seconds for the first, roughly 80 times faster, with byte-identical bodies. Semantic-gated prompt compression strips filler before forwarding, and provider keys stay server-side: the gateway stores them in its model registry and injects them on forward, so agents that talk directly to the LLM never hold them — shrinking the blast radius if an agent is compromised.

Governance teams get rolling risk scores per user, agent, and IP with automatic blocking, shadow-AI discovery for unregistered models, an M365 Copilot admin bridge with agent sync, and audit events for every configuration change — all in a single live dashboard.

Adoption requires no agent-side changes: teams point any OpenAI-compatible or Anthropic client at the gateway — including OpenAI SDKs, Claude Code, Cursor, MCP tools, or plain HTTP — and the pipeline takes over.

The Community Edition allowance is enforced by the gateway itself and re-read hourly from AI-FW's public license documents. Under the 100-million-token cap, requests flow normally; overages trigger a 7-day grace period flagged on the dashboard and in response headers before quota enforcement begins. There is no license key, no account, and no phone-home.

The AI-FW Community Edition is free to use and not open source; it is proprietary software with a published free-use allowance. Enterprise editions add extended capacity and support.

About Securetron
Securetron builds security infrastructure for the AI era. AI-FW is its AI firewall and governance gateway, sitting between AI agents and LLM backends to inspect prompts and responses, enforce policy, route models, and audit everything. Its PKI Trust Manager automates certificate lifecycle management, with Boltron-AI bringing AI-driven onboarding to PKI operations.

Press Contact:
Sarah Reid
Securetron
1 Yonge St, Toronto, ON M5E 1E6, Canada
+1 (438) 927-8036
[email protected]
https://aifw.io