Skip to main content

Choose Your Path

Pick the row that matches your goal. Each path has its own reading order and first commands.

Quick-reference decision table​

GoalTime investmentStart with
See it work in 5 minutesMinimalPath 1 — First run
Repeated local use, warm cacheLow–mediumPath 2 — Local daemon
Remote runtime with authMediumPath 3 — Remote auth
MCP integrationsMediumPath 4 — MCP
Enterprise evaluationHighPath 5 — Enterprise

1. First successful run​

Choose this if: You want the smallest possible setup and a working command tree in minutes.

What you get: A generated CLI from a sample OpenAPI file, running in embedded mode (no daemon).

Reading order:

  1. Installation — build the two binaries
  2. Quickstart — create a config, run catalog list, inspect the generated command tree

First commands to run:

./bin/ocli --embedded --config ./.cli.json catalog list --format pretty
./bin/ocli --embedded --config ./.cli.json tool schema tickets:listTickets --format pretty

Read this next: CLI overview to understand the full command model.

2. Reusable local daemon​

Choose this if: You expect repeated CLI use, a warmed catalog cache, or multiple commands against the same config.

What you get: oclird running as a local control plane; ocli points at it by URL.

Reading order:

  1. Complete Path 1 first
  2. Runtime overview
  3. Deployment models
  4. Operations overview

Read this next: Tracing and instances to understand runtime.json instance resolution.

3. Remote runtime auth and scoped access​

Choose this if: The runtime is hosted separately and access must be authenticated (bearer tokens, scoped catalogs, browser login).

What you get: Runtime bearer auth, catalog filtering, and Authentik-based login working together.

Reading order:

  1. Complete Path 2 first
  2. Security overview
  3. Authentik reference proof

Read this next: Fleet validation to see which auth paths are CI-reproducible.

4. MCP integrations​

Choose this if: You care about MCP stdio or streamable HTTP servers connecting to oclird.

What you get: Understanding of which MCP paths are available, which are proven in CI, and how to configure them.

Reading order:

  1. Complete Path 2 first
  2. Discovery & Catalog overview
  3. Deployment models
  4. Fleet validation — especially important here, as it shows which MCP paths are CI-reproducible vs. needing live proof

Read this next: Operations overview for runtime observability alongside MCP.

5. Enterprise readiness evaluation​

Choose this if: You need a review path you can hand to operators, security reviewers, or buyers — covering safe deployment, proof of functionality, and audit trails.

What you get: A structured path through real auth, runtime, validation, and policy evidence.

Reading order:

  1. Enterprise readiness — capabilities overview for evaluators
  2. Authentik reference proof — real auth flow evidence
  3. Fleet validation — CI-reproducible end-to-end proof
  4. Security overview — policy, secrets, and audit logging

Read this next: Deployment models for the deployment architectures available.