CognitiveXCOGNITIVEX · FOR ORGANIZATIONS
CognitiveX for teams & enterprises
Run the Large Cognition Model for your whole organization. One shared, governed memory with strict isolation, team seats, org-scoped API keys, and role-based access, so what one person or agent learns can compound across the company instead of resetting at every session.
WHY ORG MEMORY MATTERS
Your organization already has memory. It just leaks.
Context lives in people, threads, and tools. When an engineer leaves, an agent resets, or a project closes, the reasoning behind decisions goes with it. Most teams paper over that loss with documentation no one reads and chat history no one searches.
CognitiveX gives an organization a single living memory that every person and every agent can write to and recall from. Decisions, fixes, customer context, and hard-won preferences accumulate instead of evaporating. The Large Cognition Model runs the loop: memory, recall, reasoning, learning, evolution. The longer your team uses it, the sharper it gets, because it is keeping what it learns rather than forgetting it after each conversation.
WHAT SHIPS TODAY
A governed memory layer for the whole org.
Shared org memory, isolated
- One organization memory that members and agents read and write through the same cogx SDK, MCP, and HTTP API.
- Strict tenant isolation: your organization's memory is scoped to your org and is not shared with other tenants. See security.
Seats & roles
- Invite your team as seats under one organization. Each member works against the shared memory with their own login.
- Role-based access separates owners and members, so who can manage billing, keys, and membership is explicit rather than ambient.
Org API keys
- Issue org-scoped API keys so your services and agents authenticate against the organization, not a personal account.
- Wire a whole fleet to one memory: what one agent learns, every other agent on the org can recall through the same backend.
One org memory, three doors → cogx SDK · MCP server · HTTP API · all scoped to your organization
THE LCM PLATFORM, FOR A FLEET
Same platform. More minds writing to it.
Everything on the developer platform applies at the organization level. You are not standing up a vector database, an embedding pipeline, or a retrieval ranker. You write memories in, ask questions out, and the cognitive layer handles embedding, ranking, consolidation, and recall across every session and every member.
- Four memory tiers (
semantic,episodic,procedural,foundational) so org facts, events, how-tos, and identity are stored and recalled distinctly. - Dream consolidation compresses and links memories overnight; salience and pattern detection surface what matters across the whole org, not just one user.
- Reflection and introspection let agents reason about the organization's memory and state, not just the current prompt.
SECURITY POSTURE
Built so org memory stays org memory.
Memory is the asset, so isolation and access control are first-order concerns, not afterthoughts. Today, organization memory is tenant-isolated, access is role-based, and org API keys scope service access to the organization. Authentication uses signed JWT sessions, and traffic is served over TLS.
We are direct about what is shipped versus planned. The table below is the current picture. For the full write-up, see the security page.
| Control | Status | Notes |
|---|---|---|
| Tenant isolation | Available | Org memory is scoped per organization and not shared across tenants. |
| Role-based access | Available | Owner and member roles govern billing, keys, and membership. |
| Org-scoped API keys | Available | Issue and revoke keys that authenticate against the organization. |
| Encryption in transit | Available | TLS on all API and app traffic. |
| SSO (SAML / OIDC) | Roadmap | On the roadmap for organizations. Talk to us if it gates a rollout. |
| Audit logs | Roadmap | Org-level audit trails are planned. We will be honest about timing. |
We do not claim certifications we do not hold. SSO and audit logging are stated as roadmap, not as shipped features. If a specific control is a hard requirement for your team, raise it and we will tell you exactly where it stands.
PRICING FOR TEAMS
Priced per seat, scaled by recall.
Organizations are billed per seat. Add members as your team grows, and remove them when people move on. Writing memories is always free; recall spends credits scaled by how deep you ask the system to think, so cost tracks real usage rather than headcount alone.
- Per-seat billing for the organization, with owner-managed membership.
- Recall credits and pay-as-you-go metering carry over from the standard plans. See pricing for current rates and allotments.
- Larger or regulated deployments: talk to us about your isolation, access, and rollout needs so we can size it honestly.
QUESTIONS
Before you roll it out
Is org memory shared with other companies?
No. Each organization's memory is tenant-isolated and scoped to that organization. It is not pooled with or visible to other tenants.
Do you support SSO and audit logging today?
Both are on the roadmap rather than shipped. Today we support signed-session authentication, role-based access, and org-scoped API keys. If SSO or audit logs gate your rollout, tell us and we will give you a straight answer on timing.
How do seats and roles work?
You invite team members as seats under one organization. Owners manage billing, API keys, and membership; members work against the shared org memory with their own logins. See pricing for per-seat rates.
How is this different from giving everyone the SDK?
Individual accounts each keep their own memory. An organization unifies them into one governed, isolated memory with shared recall, role-based access, and org-scoped keys, so the company compounds its knowledge instead of fragmenting it across personal accounts. The underlying surfaces are the same platform.
Give your organization a memory.
Start building on the LCM today, or talk to us about an organization rollout, seats, and the isolation and access posture your team needs.