
At SAI Technology, we build bridges—between services, data sources, and AI models. The SAI-MCP Server is our answer to fractured APIs and brittle integrations. Powered by the Model Context Protocol (MCP), it presents a single, unified interface to Slack, GitHub, Notion, Google Sheets, Telegram, Coolify, and more.
We’re excited to share that upcoming integrations will include Zoho Books for streamlined accounting workflows and Paystack API for seamless payment processing.
Why We Built SAI-MCP
Managing a suite of tools—each with its own authentication, rate limits, and data formats—was slowing down our AI-driven workflows. We needed:
Simplicity: One endpoint, one authentication model.
Extensibility: Add services without rewriting core logic.
Consistency: Predictable request/response patterns for AI agents.
Core Architecture
Protocol-First: Implements MCP’s resources (read), tools (write), and prompts (AI patterns).
Modular Services: Each integration lives in
src/services/
, wired up insrc/index.ts
.TypeScript & pnpm: Fast installs, strict types, and zero-downtime builds.
Key Integrations at a Glance
Slack
- List channels & messages
- Send notifications and alerts
GitHub
- Query repos, issues, pull requests
- Create issues and comment on PRs
Notion
- Read pages & databases
- Automate content workflows
Google Sheets
- Fetch and update spreadsheet data
- Generate reports on the fly
Telegram
- Read and send chat messages
- Power custom bot workflows
Coolify
- Deploy & manage containerized services
- Integrate CD pipelines with AI triggers
Resources, Tools & Prompts
- Resources (Read)
Expose endpoints like
slack://channels
orgithub://repos
to fetch data. - Tools (Write)
Actions such as
send-slack-message
orcreate-github-issue
allow models to execute changes. - Prompts (Patterns)
Higher-level AI interactions (e.g.,
daily-standup
,project-status-report
) live insrc/prompts/
.
Use Cases & Roadmap
Automated Standups: Pull issues, messages & sheet data into a single daily report.
Cross-Service Workflows: Create a GitHub issue, notify Slack channels, and log to Notion in one go.
Expanded Integrations: Jira, Asana, Salesforce, Zoho Books (accounting workflows), Paystack (payments & subscriptions), even on-prem databases.
UI Console: Building a dashboard to visually compose MCP calls.
Future Plans
Our vision for SAI-MCP’s next chapters includes:
Zoho Books Integration: Leverage MCP endpoints to manage invoices, expenses, and financial reporting directly within your workflows.
Paystack Payments: Enable payment processing, refunds, and subscription management through our unified API.
Ready to unify your tools with AI?
Contact Us