Skip to content
fzymgc-house Cluster Docs

Choose Miniflux over FreshRSS as the cluster feedreader

Date: 2026-06-12 Status: Accepted Decision: hl-cgb4 Deciders: Sean Brandt

Context

Section titled “Context”

The cluster needed a self-hosted RSS feedreader to replace Readwise Reader’s reading/triage role after that subscription was cancelled. Karakeep remains as the bookmark archive, not a reader. Two mature self-hosted options were evaluated: Miniflux (Go, minimalist) and FreshRSS (PHP, feature-rich). The choice shapes which sync protocols are available cluster-wide and whether saving an entry into Karakeep requires glue code.

Decision

Section titled “Decision”

Deploy Miniflux as the cluster’s feedreader. Its native Karakeep save integration (v2.2.10+), single-binary ops profile, and open sync-protocol support (Google Reader API + Fever API) outweigh FreshRSS’s richer built-in web UI.

Rationale

Section titled “Rationale”
  • Native Karakeep save (POST /api/v1/bookmarks) eliminates a glue layer that would otherwise be a standing maintenance surface.
  • A single Go binary simplifies the Deployment manifest and removes a PHP runtime dependency.
  • Google Reader API + Fever API expose standardized sync endpoints, making the native client a swappable choice rather than a deployment concern.
  • The web UI only needs to be a competent desktop triage pane; Miniflux clears that bar.

Alternatives Considered

Section titled “Alternatives Considered”

Miniflux (chosen) — Strengths: native one-click Karakeep save (zero glue); single Go binary, minimal ops; Google Reader + Fever sync APIs; swappable front-ends. Weaknesses: more spartan web UI; fewer built-in organizational features.

FreshRSS (rejected) — Strengths: richer built-in web UI; larger extension ecosystem. Weaknesses: Karakeep save is not native (requires a bridge/extension); heavier PHP + database stack; no single-binary deploy.

Consequences

Section titled “Consequences”

Positive: No extra code or sidecar for Karakeep integration; simpler Deployment manifest (no PHP-FPM); future native-client changes need no cluster manifest changes.

Negative: Less rich web UI for power-users who prefer a FreshRSS-style folder/label hierarchy; no fallback glue layer already in place if upstream Miniflux–Karakeep integration changes.

Neutral: FreshRSS remains a viable migration target; OPML export/import keeps switching low-cost.