MCP servers for popular apps
What an AI agent can do in a named app through MCP, how to connect it safely, and where it stops.
Microsoft 365 and MCP: mail, files and Teams context, with human sending
Connect Microsoft 365 context through Work IQ MCP, review Graph permissions and tenant consent, and keep email, Teams posts and invitations under human control.
GuideAWS MCP servers: documentation, costs and safe infrastructure questions
Choose an AWS MCP server for documentation or cost questions, restrict account access with IAM, and keep production changes outside the agent run.
GuideCloudflare MCP servers: inspect first, approve configuration changes
Choose a Cloudflare MCP server, connect documentation search or scoped account access, and keep DNS and firewall changes behind human approval.
GuideSentry MCP: investigate a production error and propose a fix
Connect Sentry MCP with OAuth, inspect error evidence and reproduce a failure before proposing a reviewed fix. Understand data exposure and where Tvijo fits.
GuideShopify MCP: developer docs and catalog reads, with orders out of reach
Choose the right Shopify MCP server, connect Dev MCP in VS Code, and keep catalog questions separate from customer records, checkout and payments.
GuideFigma MCP server: design context for coding agents and its limits
Connect a coding agent to Figma design context, choose remote or desktop access, check seat requirements, and review the generated UI code.
GuideLinear MCP: from issue to agent task with a clear definition of done
Connect Linear MCP safely, turn an issue into a bounded agent task, and review the diff, verification and receipt before accepting the result.
GuideSalesforce MCP: read accounts and opportunities without widening access
Connect an agent to Salesforce’s read-only MCP server, test sandbox access, and keep CRM analysis within existing org permissions.
GuideHubSpot MCP: CRM answers and changes that need human review
Connect HubSpot CRM to Claude or an MCP client, ask grounded sales and support questions, and keep record changes behind human approval.
GuideStripe MCP: payment questions an agent can answer, and actions it must never take alone
Connect Stripe MCP for explanations and bounded reads. Start in a sandbox, restrict access, and keep refunds, billing changes and money movement under human control.
GuidePlaywright MCP: let an agent test your web app in a real browser
Connect Playwright MCP to an editor agent, test a staging workflow, and retain evidence of clicks and assertions. Learn safe setup, headless options, and limits.
GuideDatabase MCP for PostgreSQL and Supabase: SQL you can review
Connect PostgreSQL or Supabase through MCP with read-only access, inspect the SQL behind analytics answers, and understand where Tvijo fits.
GuideGoogle Drive and Workspace MCP: document access and email gates
Compare official and community Workspace MCP servers, understand OAuth scopes, and use document context while keeping outbound email behind explicit approval.
GuideJira and Confluence MCP: turn an issue into a scoped coding task
Connect Atlassian MCP safely, turn an issue and its documentation into a task packet, and review the code, checks and receipt before writing back.
GuideSlack MCP: channel summaries and drafts, with sending under human control
Connect an AI client to Slack with scoped reading access. Prepare evidence-backed summaries and drafts, then review and send messages yourself.
GuideNotion MCP: search and update pages with controlled access
Connect Notion MCP to your editor, understand hosted and local permissions, and review page changes before applying them.
GuideGitHub MCP server: what an AI agent can do in your repositories and how to limit it
Connect GitHub MCP to an editor, inspect its tools and start read-only. Understand OAuth, PAT permissions and the limits of a proposed Tvijo workflow.