Map / Runtime & Guardrails / Polaxis

Polaxis

India · Founded 2026

Quick answer: Polaxis is an AI security company based in India and founded in 2026, working in Runtime & Guardrails, Observability & Governance, MCP & LLM Gateways and 1 more category.

Runtime & GuardrailsObservability & GovernanceMCP & LLM GatewaysAI Identity

Capabilities

AI agent security and governance platform that intercepts every tool call through a 7-layer firewall — regex scan, risk scorer, LLM gate, behavioral baseline, session graph, threat intel, and policy engine; blocks prompt injection, PII leakage, credential theft, and unsafe actions before any tool executes; human-in-the-loop approvals routing high-risk actions to Slack with PII auto-masked; budget controls per agent/day/month auto-blocking when thresholds hit; immutable encrypted audit trail of every tool call, parameter, and approval; on-demand SOC 2, GDPR, HIPAA, EU AI Act, and OWASP Agentic Top 10 compliance reports; behavioral anomaly detection via Welford online stats per-agent; session-graph kill-chain detection across multi-turn agent runs; budget caps and spend visibility.

Sensors / integration

Python SDK (two-line integration) deployable in any Python environment and any agent framework; zero-code MCP proxy mode (three env vars, point MCP client at proxy); framework-agnostic — LangChain, LangGraph, CrewAI, PydanticAI, AutoGen, OpenAI function-calling, Claude SDK, Cursor, Claude Desktop, custom agents; self-hosted in customer VPC on Enterprise plans (FastAPI + MongoDB + MCP proxy).

Protections

Inline block of prompt injection, PII exfiltration, credential theft at < 1ms median Python-layer latency; 89% of calls resolved without an LLM round-trip via pure Python signals; LLM gate fires only ~11% of calls using claude-3-5-haiku with 5-min cache and fail-open; tool-call-level governance (block / allow / HITL) by tool name, params, agent identity, time, spend, or custom logic; configurable fail-open vs fail-closed per agent; tamper-proof AES-256-encrypted audit log.

Differentiator

Explicitly positioned as the tool-call governance layer — distinct from LLM-text guardrail vendors (Prompt Security, Lakera Guard) which inspect prompts and outputs; The 7-layer architecture combines deterministic Python signals with selective LLM gating to keep latency in microseconds; Free tier (1 agent, 10K evals/mo) with self-serve onboarding lowers adoption friction; Built-in compliance report generation (SOC 2, GDPR, HIPAA, EU AI Act, OWASP) for audit-ready PDFs on demand; Budget enforcement is a unique primitive — most competitors lack hard cost limits per agent; Transparent pricing ($0/$149/$499/Custom) versus traditional enterprise-only AI security vendors.

Other Runtime & Guardrails companies

AariaSec · ActaClad · ActiveFence · Acuvity · AGAT Software · AgentFox · AI Aramid · AI Prompt Shield · Aiceberg · Aim Security · AIR · Aira Security · See all →

Last updated · Get the “Listed on” badge