Learning Nodes
Learning Nodes is a decision management platform that captures the reasoning, options, and evidence behind organizational decisions from emails, Slack messages, and notes. The platform structures this information into searchable decision records, automatically surfaces relevant past decisions, and provides an open-source MCP server called Elen that gives AI coding agents persistent decision memory.
- Enterprise Software
- Artificial Intelligence
- Data & Analytics
- Developer Tools
- Software Only
- AI Agents
Funding
Founders
Founder details are not available yet.
Product
Problem
Organizations make decisions continuously, but the reasoning, questions, and evidence behind those decisions are rarely captured in a structured way. Teams start from zero every quarter, re-litigating the same choices without access to prior context, and AI agents that reason for hours forget everything once a session ends.
Solution
Learning Nodes is a decision intelligence platform that turns plain-language descriptions of situations into structured decision records. Users describe what is going on, and the platform returns the decision being faced, the questions standing in the way, and the claims that need checking, with every number traceable to its source query and file. When a decision is committed, the record is frozen and becomes searchable context for future situations. The platform also includes Elen, an open-source MCP server that gives AI coding agents persistent decision memory, allowing them to query, expand, and commit decision records using a lightweight pointer-based protocol that minimizes token overhead.
Target Audience
Primary users are business teams and decision-makers who need to document and reuse organizational reasoning, as well as AI agent developers and engineering teams who want persistent decision memory for their coding agents.
Features
- Plain-language input that automatically identifies the decision, options, and open questions without requiring forms or frameworks
- Evidence traceability showing the exact query and source file behind every computed number
- Immutable decision records that are frozen on commit, with corrections stored as new records pointing to the old ones
- Automatic surfacing of prior decisions when a new situation touches them, with cited evidence and reasons
- Elen MCP server for AI agents using a Suggest → Expand → Commit → Supersede protocol with ~200 tokens per decision
- Explicit graph structure with hash pointers and lazy loading to minimize context window usage
- Works with Claude Code, Cursor, Windsurf, VS Code Copilot, and any MCP-compatible agent
- AGPL-3.0 open-source SDK and free local MCP server