Question 1

What is gutt?

Accepted Answer

Every time you start a new AI chat, you lose all previous context. You re-explain your tech stack, coding standards, and project decisions - over and over. gutt is an AI organizational memory platform that eliminates this problem by maintaining a persistent knowledge graph that any MCP-compatible AI agent can access instantly. Your AI tools remember everything your team has taught them.

Question 2

How does gutt work?

Accepted Answer

gutt works through the MCP (Model Context Protocol). When you correct an AI agent or make a decision, gutt captures that knowledge into a graph database with temporal awareness. Future AI interactions automatically retrieve relevant context, so agents behave like experienced team members who already know your systems and preferences.

Question 3

Is there a trial or demo available?

Accepted Answer

Yes. We offer personalized demos using your actual use cases to show real impact. Schedule your demo at https://gutt.pro/contact/demo or email hello@gutt.pro.

Question 4

What AI agents does gutt work with?

Accepted Answer

gutt works with MCP-compatible AI tools. Currently this includes Claude (via Claude Code) and Cursor IDE. As the MCP ecosystem grows, gutt automatically becomes compatible with new tools.

Question 5

What data sources does gutt index?

Accepted Answer

gutt currently indexes organizational context from meeting transcripts (Google Meet), Jira issues, Confluence documentation, GitHub repositories, MS Teams conversations, and Google Drive files. Additional data sources are being added regularly - contact us to discuss your specific needs.

Question 6

What is the MCP protocol?

Accepted Answer

MCP (Model Context Protocol) is an open standard developed by Anthropic for connecting AI applications to external data sources. It enables standardized context sharing across AI agents without custom integrations for each tool. Major adopters include Anthropic's Claude and Cursor.

Question 7

How much does gutt cost?

Accepted Answer

gutt costs EUR90 per contributor per month. This includes unlimited queries, the full knowledge graph, MCP protocol integration, and team-wide learning. LLM costs are billed separately based on usage.

Question 8

What results can I expect from using gutt?

Accepted Answer

In our internal validation study with 9 engineers: 48% average time savings across tasks, 65% cost reduction on AI-assisted work, and correction rounds reduced from 3 to 1. Teams report 78% faster task completion and 54% fewer iterations by eliminating repeated context-setting.

Question 9

Is gutt suitable for small teams?

Accepted Answer

Absolutely. Small teams often see the fastest ROI because shared context has immediate impact. A 5-person team where everyone's AI agents share organizational knowledge can outperform larger teams with siloed AI usage.

Question 10

How does gutt protect sensitive organizational data?

Accepted Answer

gutt uses isolated tenant environments per organization, user-specific encryption keys, and attribute-based access control through Permit.io. Authentication is handled via Google OAuth. Contact hello@gutt.pro for detailed security documentation.

Question 11

How does gutt learn from corrections?

Accepted Answer

When you correct an AI agent (for example, "we use pnpm, not npm"), gutt captures that correction as a fact in the knowledge graph with temporal metadata. This creates a closed feedback loop where lessons are generated from corrections and past decisions, enabling AI agents to avoid repeating mistakes.

Question 12

How is gutt different from RAG or vector databases?

Accepted Answer

Traditional RAG retrieves documents but lacks understanding of relationships and how context evolves. gutt uses a Graph RAG architecture that captures not just information but organizational logic: why decisions were made, what corrections have been applied, and temporal context. AI agents understand your organization's reasoning, not just retrieve documents.

Question 13

Why not just use ChatGPT or Claude custom instructions?

Accepted Answer

Custom instructions are limited in size and don't update automatically. They're also siloed to one tool and one user. gutt maintains a comprehensive knowledge graph that works across all MCP-compatible AI tools, learns from every interaction across your entire team, and addresses the context window limitations that constrain custom instructions.

Frequently Asked Questions

Still have questions?