How It Works
gutt captures organizational understanding from meetings and tools.
You get instant answers. Your AI agents stop hallucinating.
Same brain. Both query it.
The Dual Problem gutt Solves
YOU WASTE TIME SEARCHING
Every question takes 25 minutes:
- → Search Slack threads
- → Check Jira tickets
- → Read Confluence docs
- → Ask in channels, wait for answers
You find conflicting information.
Still not sure what's correct.
YOUR AI AGENTS HALLUCINATE
When you prompt AI to generate code:
- → Uses generic patterns from training data
- → Doesn't know YOUR architectural decisions
- → Looks correct but violates your standards
- → Creates production issues
Same root cause: Missing organizational context.
gutt: ONE SOURCE OF TRUTH
gutt captures understanding from meetings and connects it to your tools. When you ask questions, you get instant answers traced to source. When your AI agents generate, they query gutt first - grounded in YOUR decisions, not generic patterns.
Same organizational brain. You both use it.
Three Steps to Build Your Organizational Brain
1. Capture from Meetings
gutt transcribes your Google Meet and Microsoft Teams meetings automatically.
Why meetings? Because that's where understanding exists at 100%.
When your team discusses architecture in a meeting, that's when everyone understands the "why" behind decisions. gutt captures it at the source - before it decays into outdated documentation or gets lost completely.
Example: In your March architecture meeting, the team decided to use connection multiplexing instead of pooling. Someone mentioned pooling caused issues before. gutt captures the decision, the reasoning, and who was involved.
2. Connect Your Tools
gutt integrates with the tools you already use:
- • Communication: Slack, Teams, Zoom, Google Meet
- • Development: GitHub, GitLab, Bitbucket
- • Project Management: Jira, Linear, Asana
- • Documentation: Confluence, Notion, Google Docs
We automatically link everything together.
Example: The March meeting decision → connected to the Jira ticket where it was tracked → connected to the GitHub PR where it was implemented → connected to Slack discussions about it.
Everything connected. Nothing disconnected.
3. Build the Knowledge Graph
gutt uses a hybrid approach: knowledge graphs combined with vector search.
We track three dimensions:
⏱️ WHEN decisions were made
Track the timeline of organizational knowledge
🔗 WHY decisions happened
Understand the relationships between cause and effect
👥 WHO was involved
Connect people, projects, and knowledge
Result:
Every answer can be traced back to the source decision.
No hallucinations. No guessing. Just truth.
Same Brain, Two Access Points
You Ask Questions
When you need to know something:
"How do we handle Redis connections?"
gutt responds instantly:
- March 2024 architecture meeting decision
- Explanation: Multiplexing chosen over pooling
- Reason: Pooling caused production issues
- Link to implementation PR
- Current standard for the team
Time: 30 seconds (not 25 minutes)
Result: Correct answer with full source traceability
Your AI Agents Query First
When you prompt AI to generate code:
"Generate Redis connection code"
Your AI agent (Cursor, custom agents) queries gutt first via MCP protocol:
- Retrieves YOUR validated multiplexing pattern
- Sees pooling marked as failed approach
- Gets organizational context
- Generates using YOUR standards
Time: Same speed, but grounded
Result: Code that works with YOUR architecture
Shared Learning from Failed Patterns
When your AI agents make mistakes, you validate:
"This approach failed"
- → gutt stores it as "don't repeat"
- → Includes why it failed
- → Includes who validated it
"This worked perfectly"
- → gutt stores it as "validated pattern"
- → Future reference for you and agents
The brain learns from every interaction.
Failed patterns are never repeated.
Both you and your agents benefit.
How AI Agents Share the Brain
Central Intelligence
Agent memory persists in the organizational knowledge graph.
Any agent - Cursor, your custom agents, future tools - shares the same central brain. They all query the same organizational understanding.
This means when one agent learns something, all agents benefit. When you validate a pattern, it's available to every agent immediately.
Collective Learning
Agents learn from collective successes and failures, creating a continuously improving system.
Every validation you provide:
→ Becomes organizational knowledge
→ Prevents future mistakes
→ Improves every agent's output
The more you use gutt, the smarter it gets.
Hybrid Approach for Accuracy
Knowledge graphs + vector search working together
Temporal Indexing
Track when decisions were made and understand the timeline of organizational knowledge.
Know what came before, what came after, and what's current.
Causal Mapping
Understand why decisions happened and the relationships between cause and effect.
Every answer explains the reasoning, not just the result.
Relationship Mapping
Connect people, projects, and knowledge for complete organizational understanding.
Know who decided, who implemented, and who can help.
Works With Your Existing Tools
No replacement. No migration. Just connect what you already use.
Communication
Slack, Teams, Zoom, Google Meet
Development
GitHub, GitLab, Bitbucket
Project Management
Jira, Linear, Asana
Documentation
Confluence, Notion, Google Docs
More integrations added regularly
Plus: MCP protocol support means gutt works with 100+ additional tools that support the standard.