Meridian

API reliability, made actionable.

Comment

What it does

Meridian is a self-hosted API reliability command center. It continuously probes critical REST and GraphQL APIs, explains failures in plain language, and turns them into actionable incidents, alerts, SLOs, and release decisions.

The problem

Teams often learn an API is broken only after a customer reports it. Traditional API clients are excellent for manually sending requests, but they do not continuously verify production behavior, connect failures to deployments, or make reliability visible to the whole team.

The solution

Meridian makes API reliability an ongoing operational workflow:

  • Create a target for a REST or GraphQL endpoint and define its expected behavior.
  • Run four clear validation layers: connectivity, authentication, contract, and data assertions.
  • Schedule recurring probes or execute them from a private-network agent.
  • Investigate recent runs, latency trends, L1-L4 failure breakdowns, and target-level analytics.
  • Create and resolve incidents, associate failures with recent deployments, and send Slack, Microsoft Teams, or SMTP alerts.
  • Define SLOs and error budgets, configure release gates, and automate multi-step transaction workflows.
  • Use role-based access, scoped CI tokens, MFA, OIDC, SSRF protection, audit history, sessions, retention controls, and alert delivery history.

The dashboard is deliberately written for people with minimal API experience: failures say what broke, where it broke, and what to try next.

Why it is different

Meridian complements tools such as Postman rather than replacing them. Postman helps a developer make a request; Meridian continuously proves that important API journeys still work after deployments, across time, and inside private networks.

Built with Codex

Meridian was built iteratively with Codex in ChatGPT using GPT-5.6 as a collaborative engineering partner. Codex helped shape the product, implement the Go API and React dashboard, tighten security, validate Docker builds and tests, and refine the experience through repeated feedback cycles.

Technology

  • Go and Gin for the API
  • React, TypeScript, Vite, and Recharts for the dashboard
  • PostgreSQL and MongoDB for persistence
  • Docker Compose for a one-command local deployment
  • Private Docker agent for internal-network probes

Get started

Clone the repository, copy .env.example to .env, set secure secrets, then run docker compose up --build. Open the dashboard at http://localhost:5173. The README includes a quick-start target, probe configuration examples, security configuration, test instructions, and a demo flow.

Built With

Share this project:

Updates