BETA FOR TEAMS SECURITY-FIRST

A squad of AI agents
that lives in your team's chat.

SquadClaw deploys agents straight into Matrix, Slack, and Telegram — they answer, retrieve, and run tools where your team already works. Built for teams. Secure by default.

Get early access → See how it works
Starts at $5/mo All channels included SSO & SCIM ready
Lives where you work
Matrix Slack Telegram Web widget API
Two pillars

Built for teams. Secure by default.

01 · For teams

Agents that live in the room with everyone.

SquadClaw isn't a single-player chatbot. Agents see the same threads your team sees, post in shared rooms, and keep context across conversations.

  • Shared agents across rooms, channels & squads
  • Per-room permissions and tool scopes
  • Threaded replies — agents respect the thread
  • Audit log: every reply, every tool call
02 · Security-first

Your data stays in your perimeter.

SSO, SCIM, scoped tokens, and per-tenant isolation from day one. Encryption in transit and at rest. Bring-your-own keys for every model provider.

  • SSO (SAML / OIDC) and SCIM provisioning
  • Per-tenant data isolation, scoped tokens
  • BYO keys for OpenAI, Anthropic, OpenRouter
  • End-to-end encrypted on Matrix rooms
Features

Everything a real squad needs.

Not just a wrapper around an LLM. SquadClaw is the orchestration layer that turns a model into an agent your team can actually trust.

Multi-channel routing

One squad, many surfaces. Route the same agent to Matrix, Slack, Telegram, or a web widget — no duplicated logic.

Tool calling that works

Hook your APIs, databases, and internal tools. Tools are versioned, scoped per agent, and rate-limited per channel.

Retrieval & memory

Drop in docs and exports. Agents recall the right snippet — and remember what was said yesterday.

Bring your own model

OpenAI, Anthropic, OpenRouter, or your own endpoint. Per-agent model selection. Costs broken out per channel and per run.

Acts on its own

Cron-driven check-ins. Event-triggered alerts. The agent reaches out without being asked — and stops being the bottleneck.

Run history & eval

Every reply gets a trace: tools called, tokens spent, latency, and a thumbs-up/down loop you can replay.

Channels

Three ways to deploy. Same squad.

Drop an agent into wherever your team is loudest. Same prompts, tools, and audit trail across every channel.

Matrix

#ops:intanova.ai
@claw what's the p99 on checkout?
last 24h → p99 412ms. Spike at 14:02 UTC tied to deploy a3f1e.

End-to-end encrypted, federated, and yours to host. The native home for SquadClaw.

E2EE FEDERATED THREADS

Slack

#support · acme
/claw refund order #8821
Found order #8821 — $24.00 · Stripe.
Refund? [Yes] [No]

Slash commands, thread replies, and approval-gated tool calls — without yet another bot to babysit.

SSO SLASH APPROVALS

Telegram

@squadclaw_bot
/status europe
All systems green ✓
EU-WEST · 41 ms · 0 incidents (24h)

For ops on the move. Group chats, DMs, and inline keyboards — perfect for paging an on-call squad.

GROUPS INLINE-KB WEBHOOK
Security

Built for teams that can't afford to leak.

SquadClaw was designed perimeter-first. Your data, your keys, your logs.

SSO & SCIM

Okta, Azure AD, Google. SCIM v2 for provisioning. Role-based access, scoped per squad.

Encrypted in transit & rest

TLS 1.3 everywhere. AES-256 at rest. End-to-end encryption on Matrix rooms — even we can't read them.

Your data stays put

EU and US regions, pinned. No cross-region replication. Per-tenant isolation. We never move your prompts or traces.

Audit log + traces

Every prompt, tool call, and refund button. Streaming export to S3, BigQuery, or your SIEM.

Pricing

Five-dollar Dev. Scale when you need to.

Start with your own model keys from $5/mo. Switch to Managed when you'd rather not think about it. Three tiers, two model strategies, one simple bill.

See pricing → Talk to us
FAQ

Things people keep asking us.

How is SquadClaw different from a Slack bot?
A Slack bot is a single endpoint. SquadClaw is a control plane: one squad, many surfaces. Same prompts, tools, and audit trail across Matrix, Slack, Telegram, and the web — and per-channel policies, not per-bot configs.
How fast is setup?
Six guided steps in the admin panel, about 10 minutes. Connect a model key, name the agent, hook a channel, send a message. No config files. No CLI for the people who shouldn't need one.
Which models do you support?
OpenAI, Anthropic, OpenRouter, and any OpenAI-compatible endpoint. You set the model per agent. Bring your own keys — billing goes straight to your provider.
What about data residency?
EU and US regions, pinned. We don't move data across regions, ever. Per-tenant isolation on shared infrastructure.
How much does it cost?
Dev tier is $5/mo with your own model keys, $19/mo if you'd like us to run the model for you. Team and Business scale from there. See Pricing for the full breakdown.
Who's behind SquadClaw?
We're Intanova — a small team building AI automation. SquadClaw is our flagship. contact@intanova.com if you want to chat.
Limited beta

Deploy your first squad
in 30 seconds.

No credit card. SSO included from day one.

Get early access → Talk to us