AI Developer Tools
Setoku
Self-hosted MCP knowledge server that makes any AI fluent in your company data with read-only queries and evolving context.
Setoku
What is Setoku?
Setoku is a self-hosted MCP knowledge server that provides AI assistants with read-only access to company data and a contextual understanding layer that improves over time.
Setoku vs Similar AI Tools
| Pricing Model | Free | Free | Free | Free |
| Free Credits | ||||
| Key Features |
|
|
|
|
| Pros |
|
|
|
|
| Cons |
|
|
|
|
| Best For |
|
|
|
|
How to use Setoku?
- 1Install the Claude Code plugin via `/plugin marketplace add Hedgy-Labs/setoku` and `/plugin install setoku@setoku`. Run `/setoku:onboard` from your project directory to set up the server, authenticate Claude, connect your database, and generate initial knowledge. Then ask questions in plain language using any MCP client (Claude, Codex, etc.).
Setoku Key Features
- Read-only SQL queries with row caps and timeouts
- Context tools that look up metric definitions and gotchas before querying
- Knowledge store that learns from corrections and human approval
- App publishing – turn answers into live web dashboards
- Human-in-the-loop for knowledge updates
- Self-hosted on your own infrastructure (single-tenant)
- Open source (Apache-2.0) with no added inference cost
Setoku Use Cases
- Query company databases using natural language without direct SQL access
- Build and publish live dashboards on real-time data
- Onboard new team members with an AI that understands data context
- Reconcile data across multiple sources (e.g., ticketing, sponsorship, payroll)
- Enable non-technical stakeholders to explore data safely
Setoku Pricing & Free Credits
Setoku currently operates on a Free model.
This tool is completely free to use
Setoku Pros & Cons
Pros
- Fully self-hosted, no data leaves your infrastructure
- Read-only access enforced at database level, not just query parsing
- Human approval required for knowledge changes, mitigating prompt injection
- No additional inference cost, works with any MCP client
- Open source under Apache-2.0 license
Cons
- Requires self-hosting and initial setup effort
- AI can still be confidently wrong despite guardrails
- Limited to supported connectors; custom sources need additional work
- Demo data is public, so security model depends on proper token management
What is Setoku best for?
- Development teams wanting to integrate AI with internal databases
- Organizations with data privacy requirements
- Teams that need to build live dashboards without frontend maintenance
- AI agents that require contextual understanding of metrics and definitions