OpenAnalytics Semantic Decision Engine
Fuse governed data, run any model or simulation, use LLMs inside explicit ontology contracts, explain uncertainty, and turn approved decisions into human or robotic OpenTasks.
LEAN PALANTIR-CLASS CORE
What OpenAnalytics governs
Semantic workflowsDAG stages with typed inputs, outputs, triggers, approvals, and allowed actions.
Any analytical modelRules, SQL, Python, ML, CV, streaming, Monte Carlo, physics, wargames, optimization, digital twins, and custom engines.
Decision provenanceModel version, scenario seed, ontology revision, sources, evidence, uncertainty, assumptions, and reviewer history.
Operational effectsApproved recommendations become ontology-validated OpenTasks for humans, services, sensors, or robotic agents.
Relationship intelligenceFuse multiple data sources into one explainable entity-and-relationship diagram with source-level provenance.
Notebook developmentRegister JupyterLab servers, discover notebooks, run them through agents, and promote reviewed notebooks into governed models or workflows.
MISSION PULSE
Recent analytical activity
No analytical runs yet.
EXECUTION PLANE
Control plane in the cloud; compute where the data and hardware live
OpenDatagoverned source and data products
→
OpenKnowledgetypes, identity, constraints, action rules
→
OpenAnalyticsworkflow, run, provenance, decision
→
Execution AgentsPython, GPU, simulation, Jupyter, Open WebUI
→
OpenTaskapproved human and machine effects
GOVERNED EXECUTABLES
Analytical Models
Define what a model consumes, produces, may recommend, where it executes, and how much human authority it requires.
SEMANTIC ORCHESTRATION
Workflows
Connect sources, mapping, fusion, models, review gates, decisions, tasks, and publication in an auditable DAG.
REPRODUCIBLE SCENARIOS
Models & Simulations
Version parameters, distributions, initial state, assumptions, iteration count, and random seed for Monte Carlo, physics, wargames, and custom simulations.
PROVENANCE & EXECUTION
Analytical Runs
Every run records the model, workflow, scenario, semantic contract, source IDs, ontology revision, executor, evidence, uncertainty, result, and logs.
HUMAN-MACHINE GOVERNANCE
Decisions
Review evidence, confidence, uncertainty, alternatives, risk, and action plans before creating real OpenTasks.
MULTI-SOURCE ENTITY & RELATIONSHIP INTELLIGENCE
Relationship Diagrams
Build one fused graph from ontology, databases, tracks, tasks, events, and custom typed records. Duplicate identities merge; provenance does not disappear.
Build the graph to fuse selected data sources.
JUPYTERLAB INTEGRATION
Notebook Servers & Governed Notebooks
Register JupyterLab servers without storing their tokens in the OpenAnalytics registry. Agents report health and notebook metadata; analysts can open notebooks, queue execution, and promote reviewed notebooks to models or workflows.
COMPUTE WORKSPACES
Jupyter Servers
ANALYTICAL ASSETS
Notebooks
EXECUTION & AI CONNECTORS
Integrations
Use direct trusted connectors when reachable, or a local OpenAnalytics Agent when models, GPUs, notebooks, robots, or private services live behind the firewall.
SEMANTIC LLM
CHECKINGOpen WebUI
No result yet.
DISTRIBUTED EXECUTION
OpenAnalytics Agents
Agents claim typed runs and execute only reviewed model references: Python files under an approved root, Jupyter notebooks under an approved notebook root, or configured Open WebUI calls. They report results, evidence, metrics, and health through OpenBus.
Agent protocol
POST /api/analytics/agents/registerGET /api/analytics/agents/:id/commands/nextPOST /api/analytics/agents/:id/commands/claimPOST /api/analytics/runs/:id/report