Gate Kali

Access AI Securely

One platform, two products. The browser extension cloaks PII and sensitive terms before your prompts reach ChatGPT, Claude, Gemini and other AI services — and Gate Kali Assistant is a self-hosted principal assistant that handles your briefings, meetings and inbox under strict approval-gated control.

Download extension Meet the assistant
Self-hosted On-device engine Approval-gated actions Audit logging

Two products. One promise.

Your sensitive data never leaves your control — whether you're chatting on someone else's AI app or ours.

Available now

Gate Kali Extension

A browser extension that sits between you and the AI tools you already use. Emails, phone numbers, SSNs, card numbers and custom terms are replaced with tokens before anything leaves your browser.

  • Works on ChatGPT, Claude, Gemini, Copilot, Grok, DeepSeek and more
  • Sanitizes typed prompts, pastes and file uploads
  • Optional zero-transit on-device WASM engine
Beta

Gate Kali Assistant

A self-hosted, multi-tenant AI principal assistant. It reads your inbox, calendar, documents and business systems — then briefs you, prepares meetings, triages correspondence and drafts replies. Every consequential action waits for your approval.

  • Google Gmail, Calendar & Drive via OAuth — you approve every scope
  • Daily executive briefings across email, calendar, tasks and docs
  • Meeting bots that join, record, transcribe and summarize
  • WhatsApp, Telegram, Slack, Teams, Discord & email channels
  • Multi-tenant, white-label web app with self-service signup

How the extension protects you

Gate Kali Extension sits between your browser and AI chat apps, tokenizing sensitive data in real time.

Protection

●

Automatic PII redaction

Emails, phones, SSNs, credit cards and more are replaced with tokens before they leave your browser.

●

Custom sensitive terms

Add project codenames, internal domains or any words you never want to reach an AI model.

●

File-upload & paste scanning

Attachments and pasted content are sanitized too — not just the text you type.

●

Injection warnings

Suspicious prompt-injection patterns are flagged as warnings so you stay in control of what you send.

Coverage

●

Works where you work

ChatGPT, Claude, Gemini, Perplexity, Grok, Copilot, DeepSeek, Mistral, Poe, Hugging Face, AI Studio — plus any custom site you add.

●

Protect custom sites

Add any domain in the popup and Gate Kali sanitizes prompts there too.

Visibility & control

●

Activity log & stats

A local log shows every sanitized send, paste and upload — with token counts — right in the popup.

●

Live status badge

The toolbar badge shows sidecar health at a glance and turns red when it's unreachable.

●

Self-hosted sidecar

Your prompts only ever reach the endpoint you configure — your data stays under your control.

●

Zero-transit mode

Optional on-device WASM engine sanitizes without a network round trip.

See it in action

A prompt with real PII becomes safe tokens before it is sent.

My email is alice@company.com and my SSN is 078-05-1120.
↓ gate-kali sanitizes before send
My email is [EMAIL_005] and my SSN is [US_SSN_001].

Gate Kali Assistant

Now in beta — a self-hosted, multi-tenant platform built on Clean Architecture: a model-agnostic router (OpenAI, Anthropic), a Postgres + pgvector knowledge base, and an approval engine that keeps a human in charge of every consequential action.

●

Context lenses

Seven built-in lenses — Personal, Diplomatic, Government, Security, Media, Commercial and Social Protocol — auto-detected to shape tone, priorities and tool access.

●

Executive briefings

Schedule, priority correspondence, deadlines and documents synthesized into a morning briefing — meeting prep included.

●

Meeting intelligence

Bots auto-join meetings, record audio, transcribe and summarize.

●

Every channel you use

Web chat plus WhatsApp, Telegram, Slack, Teams, Discord and email — Gmail, IMAP/SMTP, iCloud, Google & Apple calendars, Drive.

●

Approval-gated actions

Reads are free; sends, deletes and shares wait for your approve/reject — every decision audit-logged.

●

White-label multi-tenancy

Subdomain-resolved tenants with custom branding, feature flags and self-service signup.

Explore all capabilities →

"Brief me."
↓ schedule + inbox + tasks + documents, synthesized
3 meetings today · 12 unread, 2 flagged priority · 1 overdue follow-up · contract draft awaiting your approval.

Get access

The beta is live at mygate.gatekali.com. Sign up with your email to create your workspace — public signup provisions a tenant with an admin account, so you get your own isolated assistant instance with white-label branding controls.

Create your workspace →

Security & trust, by design

Both products share the same posture: your data stays inside infrastructure you control, and nothing consequential happens without a human decision.

●

Self-hosted by design

Deploy the sidecar and assistant platform in your own environment — Docker Compose, your Postgres, your keys, your network boundary.

●

Tenant-scoped isolation

Every repository query is scoped by tenant_id. Users, documents, approvals and skills are tenant-bound — no cross-tenant leakage.

●

Human-in-the-loop actions

Reads are allowed; sends, deletes and shares pause for explicit approve/reject. Policies are configurable per action.

●

Audit & observability

Every tool call and decision is audit-logged, with structured logging and OpenTelemetry tracing across the platform.

●

Credential isolation

A tool gateway mediates all provider access — agents and models never receive raw OAuth tokens or API keys.

●

Prompt-injection defense

Emails, documents, web pages and uploaded files are treated as untrusted content — tool-call authorization is independent of model output.

Full platform hardening details on the assistant page →

Get started

Three steps to protected prompts.

Download the extension zip and load it in chrome://extensions with Developer mode enabled — or install from the Chrome Web Store.

In the popup, set the sidecar endpoint to https://api.gatekali.com and paste your client token.

Browse to a supported AI chat site. The toolbar badge confirms the sidecar is connected and your prompts are sanitized on send.

Self-hosting or running a team?

Deploy your own Gate Kali sidecar, manage tenants, policies, custom detectors and telemetry from the admin dashboard.

Operator guide →