Five specialised agents in production today, plus Spiderbrain, our persistent-cognition product now live at spiderbrain.ai.
Specialized · 4 roles$99 / week per role
Ready-made Agent
StaffFleet starters. The fastest way to hire your first agent roles into a workflow, one week at a time.
Four specialised agents in one framework: copywriting, marketing, design, and HR. Each one trains on your brand, your work, your policies, and your jurisdictions, then drafts inside the first week.
- Four specialisations: copywriting, marketing, design, HR
- Tuned on your brand book and your past winners
- Voice-fidelity rubric: drafts below threshold get regenerated, not shipped
- HR ships audit-ready under EU AI Act Annex III, NYC Local Law 144, GDPR Article 22
RAG · Knowledgepriced by enterprise size
Second Brain Agent
GroundThe grounding layer for business knowledge. It is what the rest of your fleet answers from, every cite checkable.
A private RAG agent that learns your docs, code, contracts, and tickets. Every answer is paragraph-anchored. It never returns a document the asker cannot already open.
- RAG over docs, code, contracts, policies, tickets, chat
- Paragraph-anchored citations. Refuses when no source supports the answer.
- Permission inherited at query time from SharePoint, Drive, GitHub, Jira, Salesforce, Confluence
- Lives in your regulatory region. No cross-tenant training, ever.
Sales · Lead to paymentfrom $299 / week
Sales Agent
DirectionStaffs the revenue workflow, lead to cash. Our flagship example of a single process run end to end by a governed fleet.
A multi-channel sales agent that runs the full lead-to-payment journey. Enriches, scores, books, takes payment, and logs every consent it collects.
- One threaded conversation across WhatsApp, email, SMS, and web
- Firmographic enrichment plus 6sense or Bombora intent signal at scoring time
- In-thread Stripe Payment Link or Agentic Commerce Protocol checkout
- Per-channel consent ledger covering GDPR, CCPA, TCPA, India DPDP, WhatsApp Business policy
Legal · 66% accuracy · Counsel reviewedEnterprise licence
Predictive Legal Agent
GovernGoverned agent work in the highest-stakes workflow. Counsel supervises every output: the human-in-the-loop pattern, proven.
A counsel-supervised agent that drafts cited responses, flags regulatory drift, and refuses when no source supports the answer. Counsel finalises every output.
- NDA and vendor-paper triage with risk-tier routing
- Paragraph-level citations. Refuses when no source supports the answer.
- Regulatory diff watch: SEC, EU AI Act, DSA, GDPR, India DPDP, sector regulators
- 66% accuracy against a senior-counsel evaluator panel, model card per release
Observability · On callCustom scope
Server Agent
GovernThe guardian seed. It watches production systems today, and the same discipline extends to watching the fleet itself.
An agent that lives next to your kernel. It collapses alert storms into one triaged incident, proposes a root-cause hypothesis with a confidence score, and never acts without permission.
- Daemon, container, or K8s sidecar. Read-only by default.
- Collapses 100 alerts into one incident with one owner
- Root-cause hypotheses ranked by Bayesian confidence, refused below 70%
- Three permission tiers, global kill-switch within 5 seconds
Grounding layerspiderbrain.ai
Spiderbrain
GroundDeterministic, auditable context for engineering-heavy fleets. The same grounding layer we deploy under enterprise agents.
A living context graph that gives your AI agents persistent memory and structural reasoning, not just flat context. Same input, same understanding, a signed decision log. It knows what a change will break before you ship.
- Persistent memory and reasoning over your whole project
- Blast radius before you ship the change
- Works over MCP with Claude, Cursor, Continue and more
- Source stays private, free to start, Windows and macOS