Open-Source Workspace Intelligence for Software Systems
Learn Workspace Intelligence.
Learn the architecture behind Workspace Intelligence: how projects, policies, evidence, and changes become one shared understanding for developers, CI, IDEs, and AI agents.
Architecture Map/Simple view
- 1Input LayerProjects · Repositories · Dependencies · Changes
- 2Workspace IntelligenceModel · Graph · Impact · Verify
- 3Agent GroundingAGENTS.md · Skills · Context
- 4Evidence ContractsReports · Artifacts · Gates
- 5Shared ConsumersDevelopers · CI · IDEs · AI agents
Practical entry path
Start with evidence, not a blank prompt
Move from the concept to a machine-readable workspace artifact and verify its contract before any consumer relies on it. Each step is documented against the current CLI command surface.
- 01
Learn the system model
Start with the boundary between repository files and an evidence-backed model of the software system.
Understand the concept - 02
Build a workspace model
Run the canonical model command and inspect the versioned artifact produced from current workspace evidence.
Open the model command - 03
Verify the boundary
Validate contracts before downstream tools, CI, IDEs, or agents consume workspace intelligence.
Open contract verification
Concrete example
Follow a fact from command to governed artifact
A repository scan is only an input. Workspai turns observed facts into named artifacts with explicit schema versions, then exposes those artifacts to humans and agents through documented command boundaries.
| Command | Governed artifact | Schema version |
|---|---|---|
workspai workspace model --json --write | .workspai/reports/workspace-model.json | workspace-model.v1 |
workspai workspace contract verify --strict --json | .workspai/reports/workspace-contract-verify-last-run.json | workspace-contract-verify.v1 |
workspai workspace context --for-agent --json --write | .workspai/reports/workspace-context-agent.json | workspace-context.v1 |
Category boundary
Workspace Intelligence is more than retrieval
Indexing, retrieval, and memory remain valuable inputs. Workspace Intelligence adds a shared system model, evidence semantics, versioned outputs, and verification boundaries that multiple consumers can independently inspect.
| Capability | Primary question | Durable output | Trust boundary |
|---|---|---|---|
| Repository indexing | Where is matching code or text? | Search index | Source presence |
| RAG | What retrieved context may answer this prompt? | Prompt context | Retrieval relevance |
| Agent memory | What did an agent previously observe or decide? | Session or long-term memory | Recorded agent experience |
| Workspace Intelligence | What is this software system, and what evidence supports it? | Versioned model, graph, evidence, and decisions | Explicit contracts and provenance |
Knowledge map
Explore the complete system
What Is Workspace Intelligence?
Define the evidence-backed understanding layer for software systems.
Concepts
Separate Workspace Intelligence from RAG, memory, repository indexing, skills, and chat.
Software System Understanding
Learn the structure, semantics, relationships, intent, evolution, trust, and reasoning layers.
Mental Models
Reason about shared system models, evidence-backed context, workflows, and agent authority.
Architecture
Map the input layer, intelligence loop, evidence contracts, and consumers that share one workspace truth.
Workspace Model
Study the canonical model of projects, runtimes, commands, policies, contracts, and evidence.
Workspace Graph
Understand the graph of the software system: structure, runtime, ownership, change, evidence, and agents.
Evidence
Learn how Workspai treats verified, observed, inferred, stale, and unknown facts.
Contracts
Inspect every synced contract, field, artifact, producer command, and consumer boundary.
CLI Reference
Inspect every Workspai command and its contract-backed artifact relationships.
Guides
Adopt existing repos, create supported projects, generate agent context, and run verification gates.
RFCs
Track the product thinking behind Workspace Graph, Atlas, package separation, and future surfaces.
Glossary
Keep the vocabulary consistent across Workspai, RapidKit, docs, CLI, VS Code, and future cloud surfaces.
Essays
Read category-shaping positions on AI engineering, repositories, and shared software-system truth.
Research
Use durable evaluation frames and architectural requirements for Workspace Intelligence systems.