# InferenceWall

## Docs

- [Admin API: Reload, Stats, and Config Endpoints](https://docs.inferwall.com/api/admin.md): Admin-only endpoints to hot-reload config, view request statistics, and read runtime settings. Requires the admin API key (iwk_admin_ prefix).
- [Authenticate InferenceWall API Requests](https://docs.inferwall.com/api/authentication.md): InferenceWall uses API key authentication with scan and admin roles. Generate keys, pass them as Authorization: Bearer, and follow security best practices.
- [Health Check API Endpoints](https://docs.inferwall.com/api/health.md): GET /v1/health returns full server status; /v1/health/live is the liveness probe; /v1/health/ready is the readiness probe for Kubernetes and load balancers.
- [InferenceWall REST API Overview](https://docs.inferwall.com/api/overview.md): The InferenceWall REST API exposes scan, analyze, session, signature, policy, health, and admin endpoints — all returning JSON over HTTP at port 8000.
- [Policies API: Inspect Active Detection Policies](https://docs.inferwall.com/api/policies.md): GET /v1/policies returns all loaded policy profiles with their enforcement mode, scoring thresholds, and per-signature overrides. Requires a scan or admin key.
- [Scan Endpoints: Detect Threats in LLM Traffic](https://docs.inferwall.com/api/scan.md): POST /v1/scan/input and /v1/scan/output return allow/flag/block verdicts. The /v1/analyze/* variants include deeper PII detection detail in each response.
- [Session Management API Endpoints](https://docs.inferwall.com/api/sessions.md): Create and manage InferenceWall sessions to track multi-turn conversation context, improving detection accuracy for attacks that span multiple turns.
- [Signatures API: List and Inspect Loaded Rules](https://docs.inferwall.com/api/signatures.md): List all loaded InferenceWall signatures or retrieve one by ID. Use these endpoints to verify custom signatures are active and inspect per-signature metadata.
- [How InferenceWall Detects Threats](https://docs.inferwall.com/concepts/how-it-works.md): InferenceWall's four-layer pipeline — heuristic, ML classifier, semantic, and LLM-judge — combines results into one anomaly score driving every scan decision.
- [Policy Profiles: Configure Detection Behavior](https://docs.inferwall.com/concepts/policies.md): Policy profiles control InferenceWall's enforcement mode, scoring thresholds, and per-signature overrides — configured in YAML with no code changes required.
- [Anomaly Scoring and Decisions Explained](https://docs.inferwall.com/concepts/scoring.md): How InferenceWall anomaly scoring works: confidence times severity, diminishing corroboration, and inbound/outbound thresholds producing allow, flag, or block.
- [Detection Signatures in InferenceWall](https://docs.inferwall.com/concepts/signatures.md): Reference for InferenceWall's 100 detection signatures: categories, ID format, match object schema, ATLAS technique coverage, and engine-to-profile mapping.
- [Lite, Standard, and Full Deployment Profiles](https://docs.inferwall.com/deployment-profiles.md): Compare InferenceWall's Lite (sub-ms heuristic), Standard (80ms ML+semantic), and Full (2s LLM-judge) profiles to match your latency and accuracy requirements.
- [Configure Custom Detection Policies](https://docs.inferwall.com/guides/custom-policies.md): Create YAML policy files to tune InferenceWall thresholds, enforcement mode, and per-signature overrides without any code changes or package modifications.
- [Write Custom Detection Signatures for InferenceWall](https://docs.inferwall.com/guides/custom-signatures.md): Create custom YAML signatures to detect threats unique to your app: internal project names, proprietary data formats, or domain-specific attack patterns.
- [Deploy InferenceWall in Production](https://docs.inferwall.com/guides/deployment.md): Deploy InferenceWall via PyPI, Docker, or Kubernetes Helm chart, then configure API keys, TLS, Redis sessions, and health check endpoints for production use.
- [Ship InferenceWall Scan Logs to an ELK Stack](https://docs.inferwall.com/guides/observability.md): Ship InferenceWall scan decisions and audit events to an ELK stack for centralized security monitoring, SOC2/ISO27001 compliance, and false-positive tuning.
- [Protect Anthropic Claude Calls with InferenceWall](https://docs.inferwall.com/integrations/anthropic.md): Wrap Anthropic messages.create() with InferenceWall to scan prompts for injection attacks and Claude responses for PII and credential leakage.
- [InferenceWall HTTP Middleware for FastAPI](https://docs.inferwall.com/integrations/fastapi.md): Automatically scan every incoming HTTP request body and every outgoing response body with InferenceWall by registering a single FastAPI middleware function.
- [InferenceWall Middleware for LangChain](https://docs.inferwall.com/integrations/langchain.md): Add InferenceWall to LangChain using a wrapper function or InferenceWallCallback handler to scan user inputs and LLM outputs for threats automatically.
- [Protect OpenAI API Calls with InferenceWall](https://docs.inferwall.com/integrations/openai.md): Wrap openai.chat.completions.create() with InferenceWall to scan prompts for injection attacks and GPT responses for PII and credential leakage.
- [InferenceWall: AI Firewall for LLM Applications](https://docs.inferwall.com/introduction.md): Learn what InferenceWall is, how its multi-layer pipeline works, which deployment mode fits your stack, and how 100 built-in signatures map to MITRE ATLAS.
- [Get Started with InferenceWall in Minutes](https://docs.inferwall.com/quickstart.md): Install InferenceWall from PyPI, scan your first user input and LLM output for threats, choose a detection profile, and run the standalone API server.
- [InferenceWall CLI Command Reference](https://docs.inferwall.com/reference/cli.md): Complete reference for all inferwall CLI commands and flags: serve the API server, scan single inputs, download and manage ML models, and generate API keys.
- [InferenceWall Environment Variables Reference](https://docs.inferwall.com/reference/environment-variables.md): Reference for all InferenceWall env vars: API keys, detection profile, host/port, signatures directory, policy file path, ELK logging, and Redis session store.
- [InferenceWall Detection Signature Catalog](https://docs.inferwall.com/reference/signature-catalog.md): Full reference for all 100 InferenceWall signatures: MITRE ATLAS mappings, severity levels, engine types, and detection coverage across 5 threat categories.