The problem it solves
AI assistants accelerate typing, not engineering.
nablr is the orchestration layer that fixes this - quality enforcement, workflow context, and persistent state that AI assistants lack by default.
Source: SEF: A Multi-Agent Orchestration Framework for Parallel SDLC Execution - DOI 10.5281/zenodo.18410037.
When to use nablr
Reach for nablr when:AI code ships defects
You use Cursor, Copilot, or Claude but can’t trust what merges.
Requirements aren't buildable
Stories get written but QA finds gaps after development.
Incidents recur
Same root causes keep appearing with different symptoms.
Codebase is unknown
Inherited system, no docs, no safe entry point for change.
Migrations stall
Platform rewrites start without a complete map and discover unknowns mid-flight.
Governance is required
SOC2, ISO 27001, or internal audit trails needed for AI-assisted development.
How it works
nablr runs as a local MCP server alongside your AI assistant. When you invoke a persona, nablr injects the relevant rules, standards, and context directly into the conversation.Free vs paid
Compatible clients
nablr uses stdio transport and works with any MCP-compatible client:- Cursor
- Claude Desktop
- Windsurf
Gemini and GitHub Copilot do not currently support MCP.