Architecture
The singularity is here.
PromptQL provides a stable multiplayer environment with a shared context backbone so that you can continuously absorb the incredible pace of AI innovation.
Models
+
+External SaaS & data
+
+01Surface
A stable entrypoint to AI work.
Multiplayer·Cutting edge
The best of ChatGPT, Claude Code and Cowork, on any model. Teams never switch tools again.
- Multiplayer bots
- Chat, code & apps
- Web, desktop & mobile
- Slack & Teams
- API & MCP
Chat
Artifacts
VM
$
$
$
ModelClaudeGPTGeminiGrokNext
ClaudeGPTGeminiGrokNext02Shared context
Context your org owns.
Shared·Portable
Every bot adds to it. It’s yours to take anywhere, even to your own Git repo.
- Knowledge
- Skills
- Semantic layer
- Ontology
- Git export
Shared context
Knowledge
What “active” means
Skills
How we run a QBR
Semantic Layer
revenue = Σ invoices
Ontology
Account → Team → User
acme/contextmain
knowledge/active-user.md
skills/run-a-qbr.md
semantic/revenue.yml
ontology/account.md
03Infrastructure
Runs in your cloud.
Single-tenant·Your cloud
No multi-tenant app layer. Data, keys and compute stay inside your perimeter.
- Dedicated data plane
- Compute & storage
- Credentials vault
- Access policies
Your cloudAWS · GCP · Azure
Single-tenantCompute
Storage
Credentials vault
Access policies
Battle-tested storage & computemay be multi-tenant
DeployBYOCBYOKDedicated
Changes over time
Plug in what’s next.
Models ship constantly. SaaS tools come and go. Swap them in; the three layers hold.
ModelsClaudeGPTGeminiGrokDeepSeekQwenWhat’s next
ClaudeGPTGeminiGrokDeepSeekQwenWhat’s nextExternal SaaS & dataSnowflakeDatabricksPostgresSalesforceGitHubJiraNotionWhat’s next
SnowflakeDatabricksPostgresSalesforceGitHubJiraNotionWhat’s nextSurface · Shared context · Infrastructure