Kawkav agents aren't chatbots — they're configured workers, deployed on your tenant, powered by private LLMs. Pick the role; we deliver the agent.
A fast, multilingual messaging specialist — handles inbox triage, drafts replies in your voice, coordinates outreach, schedules across calendars. Hermes lives where your conversations live.
Reads, classifies, summarizes. Surfaces what needs you. Auto-archives noise. Works with Gmail, Outlook, IMAP.
Trained on your past correspondence. Tone, cadence, sign-offs. Drafts ready to send or auto-send on whitelist.
Google, Microsoft 365, CalDAV. Proposes times, books rooms, sends invites, handles reschedules and timezones.
Email sequences, Slack, Teams, Telegram, SMS. Cadence rules. Reply detection auto-pauses sequences.
English, Spanish, French, German, Catalan, Portuguese, Italian out of the box. Adds languages on request.
HubSpot, Pipedrive, Salesforce, Attio. Every conversation logged. Lead status updated. No double entry.
Hermes Pro: 500k tokens/day, unlimited inboxes, 10 users, priority queue, dedicated phone-home model option for sensitive industries.
A maker's agent — researches across the open web and your private knowledge base, executes code in a sandbox, opens GitHub PRs, builds a working prototype while you're in a meeting.
Runs Python, Node, Bash in isolated micro-VMs. File system, packages, network — all yours. Outputs returned to chat.
Search, scrape, summarize. Cites sources. Respects robots.txt. Bring-your-own SerpAPI / Brave keys or use ours.
Ingest PDFs, docs, websites, code repos. Per-tenant vector DB. Citations baked into every response.
Read repos, open PRs, comment on issues, review code. Auth via GitHub App per tenant. Fine-grained scopes.
Ollama, vLLM, LangChain, Chroma, Qdrant. Open-weight models. No vendor lock-in. Bring your own fine-tunes.
Define your own tools via OpenAPI or simple JSON schema. Open Claw picks the right one. Audit log everywhere.
Open Claw Team: 1M tokens/day, 4 hours sandbox, 1M vector chunks, 10 users with separate vector spaces per user.
A hardened agent for regulated environments — incident response, policy enforcement, access reviews, audit trail generation. Built for hospitals, insurance, government, finance, legal.
Splunk, Elastic, Wazuh, QRadar. Reads events, correlates, drafts incident reports, files tickets in Jira/ServiceNow.
First-responder playbook execution. Isolates hosts, rotates secrets, summons humans with full context. Logged end-to-end.
SOC 2, ISO 27001, HIPAA, GDPR. Continuous evidence collection. Auditor-ready reports on a quarterly schedule.
Quarterly RBAC reviews on AD, Okta, Google Workspace. Flags stale permissions. Drafts removal requests for sign-off.
Air-gapped, single-tenant, customer-controlled keys. Sovereign LLM runs entirely on your hardware if required.
Iron Claw can't change anything without an approved playbook. Every action is gated by RBAC and human sign-off.
Customer-controlled key (HSM), data residency proofs, scheduled key rotation, signed evidence chain, DPA + SCC + sub-processor list.
Tell us about your team — we'll suggest the right agent, set it up with you, and get you a working pilot inside 7 days.
Book a fit call → Back to home