CognetivyCognetivy is an open-source state layer for AI coding assistants that adds workflows, run tracking, and persistent collections - all local.
Technical Architecture Research
Research and compare architectural options with a structured, traceable process.
Architecture decisions require gathering evidence from docs, benchmarks, community discussions, and existing codebases. Cognetivy gives your agent a workflow to systematically research options, store findings in structured collections, and produce a decision brief with full source traceability.
What you'll achieve
Evidence-based decisions
Every score links to the specific docs, benchmarks, or discussions that justified it.
Persistent decision history
All ADR artifacts accumulate - future team members can understand why decisions were made.
Consistent evaluation
The same scoring criteria are applied to every option in every ADR.
How it works
A repeatable workflow you define once and run any time.
Define the decision context
A workflow step takes the problem statement and constraints and stores them as the reference record.
Research each option
Your agent reads official docs, community discussions, and benchmarks for each candidate architecture.
Score against criteria
Each option is scored against your decision criteria: performance, maintainability, ecosystem, migration cost.
Produce decision brief
A structured ADR (Architecture Decision Record) artifact is generated, with scores and source citations.
Cognetivy features you'll use
Everything you need is already in Cognetivy.
Versioned, reusable step-by-step processes for your agent
Schema-backed data stores that persist and accumulate across runs
Structured outputs saved from any run - downloadable and linkable
Local browser UI to inspect runs, events, and artifacts
Related use cases
Ready to try?
Set up your first structured workflow in under a minute.