Security Core
Built like infrastructure.
Not a chatbot wrapper.
Automated PII Redaction
Names, emails, SSNs and custom entities detected and redacted in real time — before they leave your network.
Secret Detection
API keys, database credentials and SSH keys caught mid-prompt. Block or obfuscate, your choice.
Total Audit Trail
Every interaction logged with safety scores. Know exactly what your organization sends to AI.
Team Policies
Granular redaction and access rules per workspace, team or user — managed from one dashboard.
Incognito Mode
Zero-history conversations. Nothing saved to history, nothing visible to others — privacy on top of privacy.
Model Agnostic
OpenAI, Gemini, Claude, Grok, DeepSeek, MiniMax — switch providers without changing your security posture.
Inside the Gateway
Every capability, built in.
Not add-ons. Not plugins. Everything below ships in the gateway, configured per workspace.
01 · Protection Modes
Block it, or obfuscate it.
Choose how the firewall reacts when sensitive data is detected. Per workspace, inherited by every member.
Obfuscate
PII replaced with placeholders. Prompts still work, real data never leaves.
Block
Sensitive messages stopped cold, with a clear reason shown.
Disabled
Full passthrough. Still logged, still visible in metrics.
Your message was blocked because it contains sensitive information.
Reason
Secret detection gate blocked: Aadhaar Number
Detected
Aadhaar Number
Your message was obfuscated to protect sensitive information.
Detected
Email Address
02 · Interactive Artifacts
Empower developers.
Code, documents and spreadsheets in an interactive panel beside the chat. Created, edited and versioned without leaving the conversation.
Code artifacts
Syntax highlighting, inline editing and live preview in a sandbox.
Version history
Step through every revision and roll back with one click.
Docs & sheets
Rich text with diff view, and data grids with CSV export.
03 · Teams & Controls
One dashboard. Total control.
Isolated workspaces with their own members, policies, models and budgets.
Roles & invites
Invite by email, assign member or admin, remove instantly.
Token quotas
Hard limits per workspace with real-time usage tracking.
Feature toggles
Enable or disable AI chat, uploads and sharing per team.
04 · Complete Visibility
Know what your org sends to AI.
A live Security Control Center for admins. Safety scores, violations and usage across the whole workspace.
Safety Score
Percentage of requests that passed your security gates, in real time.
Per-user breakdown
Messages, tokens, blocked incidents and uploads for every member.
CSV export
One click to a full metrics report for compliance and audits.
The Platform
One core. Everything flows through it.
Every product talks to the same firewall engine. Policies, redaction and audit enforced once, everywhere.
Your team uses
Core
Safely reaches
OpenAI
Gemini
Claude
Grok
DeepSeek
MiniMaxPlatform
The same firewall, wherever you work.
API, terminal, editor and phone — one account, one policy, every surface.
01 · API
Drop-in. Not rip-out.
A fully OpenAI-compatible API. Change one URL and every existing SDK, script and integration is firewalled.
List models
GET /v1/models
Fetch all available workspaces and models through the gateway.
Chat completion
Streaming response
Full OpenAI-compatible chat completions with real-time streaming.
Obfuscate mode
PII stripped before send
Sensitive data replaced with placeholders — prompts still work.
Block mode
Secrets intercepted
Messages containing secrets stopped before reaching the provider.
Login
Connect any provider
One command to connect to OneFirewall or any OpenAI-compatible provider.
02 · CLI Agent
AI in your terminal.
onefirewall-ai is a Python CLI that connects to your OneFirewall gateway or any OpenAI-compatible provider. Chat, stream and run agent workflows without leaving the shell.
Agent REPL
Chat and run tasks
Interactive REPL with agent mode — inspect files, run tools, and iterate.
File-aware
Reads and edits your code
The agent reads your codebase and makes real changes to your files.
Pipe & one-shot
Analyze logs and diffs
Pipe logs, diffs or files for instant AI-powered analysis.
03 · VS Code
Protected AI, in your editor.
Sign in once with a magic link, pick your workspace, and every AI request from your editor goes through the gateway.
04 · Mobile Apps
The firewall in your pocket.
Native iOS and Android apps with the same engine, the same policies and the same models as the web gateway.
Why we exist
Every day, employees paste customer data, API keys and contracts into public AI. Banning AI doesn't work. Securing it does.
Intelligence Layer
Temporal intelligence & tools.
The gateway injects precise context and high-performance tools into every conversation, ensuring real-time accuracy.
Time-aware reasoning
Current date, time and longitudinal context injected into every prompt.
Built-in toolsets
Web search, weather and custom business APIs — no model configuration.
Long-term memory
Per-user AI memory persists context across conversations and devices.
FAQ
Questions teams ask.
An enterprise AI security gateway that sits between your team and public AI providers — preventing data leaks, redacting PII and enforcing privacy policies on every prompt, in real time.