- Connection-aware Welcome: when the hub is down, show a hero with a
primary "Start hub" CTA + install fallback instead of a dead,
all-unchecked onboarding checklist (the first-run cliff).
- Actionable binary-not-found (file picker + install link, not a
"set FAI_BIN" dead end) and a connect-failure banner after
repeated failed health polls.
- Localize six hardcoded English error/toast clusters (DE+EN ARB).
- Bundle Inter + JetBrains Mono as assets; drop the runtime
google_fonts fetch (air-gap / KRITIS safe, no font-swap flash).
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
Bundles the error-UX overhaul and the inline log viewer:
- New `FaiLog` (`~/.fai/logs/studio-errors.log`, 256 KiB rotation,
JSON-per-line). Every operator-visible failure is appended so
`fai admin doctor` and the new viewer can show the trail
without the operator having to reproduce the failure.
- New `showFaiErrorSnack` / `showFaiErrorDialog` helpers wrap
`FaiErrorBox` in copyable surfaces; 27 ad-hoc
`SnackBar(content: Text(e.toString()))` sites swept to use
them (settings, doctor, audit, store, module sheet, system-AI
editor, flow output).
- New `FaiLogViewer` modal (`showFaiLogViewer`) renders log
files inline with line numbers, JSON-key + `[level]` token
colouring, Copy-all, Refresh, Open-externally. Doctor's
daemon-paths panel grows a "View" button next to "Open" for
every `.log` row and now also lists the Studio errors log.
- Today carousel: CTAs now actually re-run search after a
`filterCategory` / `runQuery` story is tapped (was only
flipping the chip state). Fallback story list bumped to 8.
- Editor bumped to git ref carrying 0.15.0 (type-token
colouring + analyzer diagnostics).
Studio bumped to 0.62.0.
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
Settings dialog's Theme Plugin section is now a grid of
swatched tiles:
- Built-in (none) — falls back to FaiTheme.light/.dark
- One tile per installed studio.theme.* plugin, each
showing the plugin's primary/secondary/tertiary as
live colour dots. Tile loads its preview lazily so a
dozen installed themes don't block the picker.
- Custom — opens a colour-picker dialog with 12 curated
Material presets + a hex input + live preview. Selecting
applies ColorScheme.fromSeed for both brightnesses.
main.dart's _pluginThemes parses a 'custom:#RRGGBB' sigil
in the same notifier slot as plugin capability ids, so the
existing persistence + restoration paths cover the custom
case with no new state.
Bumps editor to 0.11.0 (type-checked port connections +
dynamic card width fix + card-height border allowance) and
Studio to 0.58.0.
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
Two clean-ups:
- `fai_system_ai_editor.dart` uses the new `RadioGroup<T>`
ancestor pattern Material 3 introduced after Flutter 3.32.
Same UX (three radios for off/redacted/full), no
deprecation warnings, `flutter analyze` is now zero-issue
for the whole project.
- New `install_install_planned_test.dart` asserts that
installing a `planned` seed entry surfaces a clear,
human-readable error class (no panic, no silent success).
Locks in the failed-precondition path Studio's
friendly-error mapper depends on.
Integration test suite is now 3 scenarios:
- capabilities_test.dart: system.approval+kind tag
- install_install_planned_test.dart: planned-install error
All run against a fresh `fai serve` subprocess via
HubFixture, ~2 s total.
Signed-off-by: flemming-it <sf@flemming.it>
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
- Localize Flows page: app-bar title and reload tooltip,
hub-unreachable / no-saved-flows empty states, run-flow input
dialog (title with flow name, description, hint, Cancel /
Run), running dialog (title, "Flow running…" status, no-output
message, Close button), flow-card Run button.
- Localize the module-sheet bottom sheet: failed-to-load text,
Capabilities and "Declared permissions" section headers,
no-permissions placeholder copy.
- Localize the System-AI configuration dialog: title, intro
paragraph, provider dropdown label, endpoint label, API-key
env-var label (required vs optional) and disclaimer, privacy
mode header and three options (Off / Redacted / Full) plus
their descriptions, test-result panel ("Connection ok" /
"Connection failed", "Reply: …" prefix), model picker (label,
hint fallback, helper text with Ollama variant, Refresh /
Pull / Pulling… buttons, list errors and empty states),
hardware banner ("Detected: …" + " · curation reviewed …"
suffix), suitability legend (recommended / balanced / small /
large / huge / unknown), cache row with pluralized count and
Clear button, cache cleared / clear-failed toasts, pull-empty
/ pull-failed errors. Suitability label moved off a getter
onto a `labelFor(BuildContext)` method so it can read the
current locale.
Provider preset descriptions and modelHint strings stay in
English in `_ProviderPreset.all` — they're tightly coupled to
the wire-protocol identifiers and would require a runtime
factory rebuild rather than a const list.
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
Audit drilldown's explanation panel now surfaces whether the
answer came from the cache:
- "cached" pill next to the SYSTEM AI label, tooltip shows
the original generation date + hit count.
- Latency renders as "orig 47000 ms" so the operator sees
what a live call would have cost.
- Refresh icon next to the latency triggers Regenerate —
drops the entry, re-asks the model.
System-AI editor shows a cache-status row when the cache is
non-empty with a one-click Clear button. Hidden on fresh
installs to keep the dialog quiet.
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
System-AI editor now fetches the detected hardware tier and the
hub's curated model database on open. Each model chip's colour
+ tooltip combines both: green star when curated quality is
good/excellent AND min_hw_tier ≤ host; orange when quality is
fine but the host is below min_hw_tier; warning for "basic"
curated entries. Tooltip exposes the curator's note, parameter
count, context window, and licence. Banner above the form
shows "Detected: <summary> · curation reviewed <date>".
Sidebar footer gains a small monospaced live clock between the
theme toggle and the settings icon. Tooltip exposes full local
ISO + UTC for cross-checking against audit-log timestamps.
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
Two operator-felt papercuts on the editor:
* Models only appeared after a manual Refresh click. Now the
editor calls `_refreshModels()` from initState (post-frame)
so chips populate immediately. Provider switch also re-fires
the refresh because endpoint / model set changes too.
* Operator had no signal which model is suitable. Tiny models
fail at instruction following; huge models stall on a
laptop. Adds a heuristic suitability rating (parses
parameter count from common names like `gemma3:4b`,
`llama3.2:70b`; falls back to family hints for OpenAI /
Claude). Each chip now carries:
- a coloured leading dot (green=balanced, orange=small
or large, red=huge, grey=unknown)
- a star icon for the hand-curated "tested with F∆I"
allowlist (gemma3:4b, llama3.2:3b, qwen2.5-coder:7b,
gpt-4o-mini, claude-haiku-4-5, …)
- a tooltip ("balanced", "huge — likely too slow on a
laptop", …)
Sorting also moves recommended → balanced → others to the
front so the operator's eye lands on green first.
* Compact legend below the chip wrap explains the colour
code in place — no lookup needed.
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
Three operator-facing improvements driven by Stefan's testing:
* **Test Connection** runs against the form's *current* values
and no longer requires saving first. Previously we save-then-
test which caused a "system AI is disabled" message when the
form's model field was still empty. Now the test prompt
travels with the request body so unsaved tweaks can be
validated.
* **Model picker** replaces the free-text Model field. A
Refresh button calls HubAdmin.ListSystemAiModels (provider's
`/v1/models`); installed models show up as clickable chips
that fill the field with one tap. The text input stays for
cases where the operator already knows the id.
* **Pull** button (Ollama only) downloads a model via
`/api/pull`. Type `gemma3:4b`, click Pull, spinner, then the
new model lights up in the chip list. No more "switch to a
terminal to ollama-pull" detour.
Helper text updates for the empty-model case ("Required: pick
from the list (Refresh) or type one. Use Pull to download from
Ollama.") so a fresh operator never wonders why Save / Test
errors out.
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
Replaces the read-only System AI block in Settings with a real
editor. Operator never has to touch ~/.fai/config.yaml again
to change provider, endpoint, model, API-key env-var, or
privacy mode.
UX choices that align with the zero-learning-curve rule:
* Provider dropdown lists Ollama / OpenAI / LM Studio / vLLM
/ Custom by friendly name. Each selection auto-fills the
endpoint and api_key_env defaults — but only when the
field is still empty or matches a different preset's
default, so manual overrides are never clobbered.
* Each provider has a one-line description rendered under
the dropdown ("Ollama → Local `ollama serve`. Models stay
on your machine. No API key needed.") and a model-hint
placeholder ("gemma3:4b · llama3.2:3b · qwen2.5-coder:7b")
that goes away when a model is typed.
* Privacy mode is a vertical radio group with each option's
description in place — no doc-lookup needed.
* "Test connection" sends an `ok`-ping and renders the
same error-kind → fix-hint mapping the audit drill-down
uses.
* "Save" persists + hot-reloads the hub. No daemon restart
required. UI status badge flips to "enabled · <privacy>"
immediately.
`fai_dart_sdk` 0.5.0 carries the underlying `updateSystemAi` /
`testSystemAi` RPCs.
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
