Dagger
title: Dagger type: platform created: 2026-05-26 last_updated: 2026-05-26 related: ["radar/techniques/ContinuousIntegration", "radar/tools/Just", "radar/tools/Bashunit"] sources: ["https://dagger.io/"] radar_quadrant: Platforms radar_ring: Assess radar_position: inner
Portable CI/CD engine that lets developers define pipelines as code (Go, Python, or TypeScript) running in containers, with identical execution locally and inside any CI system.
Dagger pipelines are written using the Dagger SDK in a real programming language. At runtime, the Dagger Engine executes each step inside containers and caches the dependency graph — steps that did not change are skipped, whether the pipeline runs on a developer's machine or in a CI runner.
Core properties:
-
Local execution —
dagger runexecutes the full pipeline on the developer's machine with the same container graph used in CI; reproduce CI failures without pushing commits - CI-provider portability — the same pipeline definition runs on GitHub Actions, GitLab CI, CircleCI, Jenkins, or any runner that can execute a Dagger call; no provider-specific YAML required
- SDK-first — pipelines compose like functions in the host language; testable, version-controlled, and reusable across projects
- Automatic caching — container layer cache is shared between local and CI runs; incremental builds without per-provider cache configuration
Dagger addresses the pain of YAML-only CI pipelines that can only be debugged by pushing commits. It is most valuable for teams with complex multi-step pipelines, multiple CI providers to target, or a need to reproduce production builds locally. For simple projects where GitHub Actions YAML is sufficient, the SDK abstraction adds more overhead than it removes.
Dagger was founded by Solomon Hykes (Docker co-creator). The container execution model reflects that lineage — the caching and layer-sharing behaviour is well-engineered.
Dagger sits in the Assess ring of the Platforms quadrant, at inner position. First studied via dagger.io (2024-08-15). The portable CI-as-code model fills a genuine gap — no existing radar entry covers CI pipeline portability and local execution together. Inner position reflects a clear trial path (install the CLI, write one pipeline step in Python or Go, run locally) and growing adoption signal. Remaining gate before Trial is a complete pipeline running identically in local dev and a real CI environment for an active project.
