Adds an "AI parsing" section under Settings -> Integrations where a user can choose the LLM provider, model, base URL, API key and multimodal option used for booking extraction. This per-user config applies when an admin has not configured an instance-wide model. Reuses the existing encrypted user settings: the API key is stored encrypted, never prefilled, and a blank field keeps the stored one. Adds settings.aiParsing.* across all 20 locales.
The source document was only kept in memory on the background task, so a page reload during the (now always-LLM ~25s) parse lost it and the booking saved without its file. Store the uploaded files in IndexedDB keyed by job id; the review loads them from there when the in-memory copy is gone, and a 1h TTL prunes abandoned imports.
During the per-item import review the booking isn't saved yet, so the Costs section showed an empty 'Create expense' even though a linked cost will be created on save. Show the parsed price (amount + category) as the pending linked expense so the user can verify it up front. Reuses existing i18n keys.
A long single-booking PDF (e.g. an 11-page rental voucher) spent ~200s on CPU prompt-eval at the 16k cap, though its data sits in the first ~2k. Cap non-flight docs at 6k (flights keep 16k for all legs). Also make the rental operator a required field so the car gets a real title.
The saving client gets no budget:created echo (X-Socket-Id) and the create response omits the linked budget item, so the booking's Costs section and the Costs tab stayed stale until a manual reload. Reload the budget items right after a create that carried a budget entry.
After dropping the vendor templates, the model skipped the (often unlabeled) Expedia-style hotel address — making address a required schema field forces it to emit the street-address line, restoring the booking's location/place. Also hint the rental company so a car booking gets a real title instead of the generic fallback.
Now that a capable instruct model (Qwen3-8B, thinking off) reads name/address/dates/legs reliably across formats, the per-vendor template short-circuit distorted more than it fixed: brittle on layout variations and overriding the better model output. Remove the template layer; the model extracts the structure and Schicht 2 backfills the confirmation/total and takes the currency from the document's own symbol (correcting model misreads like ¥→$). Per-type prompts now also ask for address and price/currency.
A/B against the prior default (qwen2.5:7b) on CPU showed Qwen3-8B is both faster and more accurate on tricky/multilingual booking docs (correct Airbnb year+price, correct DisneySea admission date), once thinking is disabled — which the router now does. Feature it as the recommended pull, keep qwen2.5:7b as the fallback.
Hybrid/reasoning models (Qwen3 and similar) default to emitting reasoning tokens, which collide with Ollama's format-grammar constraint — on CPU this produced null/unparseable output and blew the latency budget (qwen3:8b: null or 300s timeouts vs ~20s with thinking off). Send think:false on the /api/chat call; Ollama ignores it for non-thinking models (verified on qwen2.5:7b), so it's safe and unlocks the stronger Qwen3 family.
Keep the uploaded files on the background task and hand them to the review flow, so each reviewed booking pre-fills its Files with the document it was parsed from (uploaded with the booking on save). The two modals also adopt the shared resolveDayId helper.
Code-audit clean-ups: share one normCurrency between the router and the templates, lift the duplicated nearest-day resolver into formatters.resolveDayId, drop two needless as-unknown-as casts at the fillBookingWideFields call sites, restore routeExtraction's doc comment, and give the broker template readable names. Plus recognise ¥/JPY and fall back to a standalone symbol amount, so a Klook-style voucher whose price sits far from any label still yields a cost.
Imported bookings auto-create their linked budget items server-side, but the saving client suppresses its own budget:created echo, so the Costs list stayed stale until a manual reload. Reload the budget items when the review session ends.
A reviewed transport (e.g. a rental car) arrived with only its parsed pick-up/return dates and no day_id, so the modal kept just the time and saved a bare "HH:MM" with no date. Resolve start/end day from the parsed dates (exact match, else nearest trip day) so the booking lands on the right days.
reservation_endpoints.lat/lng are NOT NULL, so saving a reviewed transport whose pick-up/return couldn't be geocoded threw a 500 and lost the whole booking (dates, linked cost). Skip those rows; the dates still persist on reservation_time/reservation_end_time.
Persist the background-import tasks (id/trip/status only) and re-fetch each job's status on mount, so a parse still running when the page reloads keeps its widget instead of vanishing; expired jobs (404) are dropped and a restored 'done' task re-fetches its items.
Apply the deterministic confirmation-code and total fill to vendor-template results too (not just model output), and require the captured reference to contain a digit so a bare 'Confirmation'/'Reference' label no longer grabs the next prose word.
Pull the hotel/rental fields these vendors print in a stable text layout (name, address, stay/pickup dates, price, reference) deterministically, so the import stops depending on the local model for them. Handles German long/abbreviated months and English dates incl. 12-hour and comma forms.
Reviewing an imported booking saves it through the normal reservation
form, which dropped the parsed price (so no linked cost was created) and
only created the accommodation when both nights matched a trip day.
Carry the parsed price into a linked cost on save, and create the
accommodation from whichever day the check-in/out dates resolve to.
Parsing a booking can take a while on a CPU host, so don't hold the
upload modal open for it. The async import endpoint returns a job id
right away; the parse runs server-side (one at a time per user) and
pushes progress over the user's WebSocket, and a small widget in the
bottom corner tracks it while the user keeps navigating and editing.
A finished job opens the per-item review from the widget.
The single-shot prompt was unreliable on multi-leg flights and longer
documents, and slow on a CPU host. For the local provider, run a small
router instead:
- deterministic vendor templates first, with no model call at all
- exactly one grammar-enforced call per document via Ollama's native
`format` (flights as a flat array of legs, everything else as one flat
reservation, the type picked from keywords or a union schema)
- booking-wide fields (booking reference, total price, the overnight
arrival day) filled deterministically from the text afterwards, and
dates coerced to ISO so a natural-language date can't slip through
Recommend qwen2.5 in the AI-parsing settings instead of NuExtract.
Instead of writing parsed items straight to the trip, the import opens the
normal edit modal pre-filled for each one, so you can check and fix it before
saving — useful when a model guesses a wrong date or address. Hotels gained an
editable address field; on save an existing place is matched by name, otherwise
the reviewed address is geocoded and a new place is created.
The provider picker is the shared CustomSelect now and the form is split into
clear sections rather than a flat stack of inputs. NuExtract 2.0 2B is the
recommended default — fastest on a CPU-only host and MIT licensed; the 4B
carries a non-commercial licence, so it's no longer flagged as recommended.
NuExtract isn't an instruct model — fed a plain chat prompt it just echoes the
schema back. Detect a NuExtract model by id and talk to it the way the model
cards document: the JSON template inlined in a single user message, no system
prompt, no json_schema, temperature 0. Its flat result is mapped back to the
same KiReservation shape the rest of the pipeline already uses, so nothing
downstream changes; every other model keeps the generic prompt.
Money is taken as a verbatim string and parsed locally (German "1.580,22 €"
otherwise comes back as 1.49772), a rental car's pickup/return ride the from/to
fields so a stray form label doesn't become the location, and a lodging with no
name falls back to its address instead of being dropped.
A freshly imported hotel links to an accommodation that lives outside the
trip store, so loadTrip alone left the reservation edit modal with blank
place/date fields. Reload the accommodations list once the import finishes.
When a confirmation carries a total price, record it as a real expense
linked to the reservation (in the matching Costs category) instead of
leaving the amount in metadata only. Gated on the Costs addon.
Request and map root-level seat/class/platform and a total price/currency into reservation metadata (shown on the card; price reuses the existing label). Read both the root and reservationFor and tolerate common field-name aliases (priceAmount, priceCurrencyISO4217Code, fareClass, ...) since models name these inconsistently. Also capture event/attraction venue telephone + url onto the auto-created place, matching lodging/restaurant.
On a GPU-less host the model's prompt-eval time scales with input length and dominates total latency. Booking details sit at the top of a confirmation, so capping the extracted text at 4000 chars (was 8000) roughly halves extraction time (~50s warm for a capable local 7B model) with no loss of fields on real hotel/rental confirmations. Tunable if a long multi-segment itinerary needs more.
- rental car: request+map dropoffLocation, emit pickup->return from/to endpoints, set a location string (G1/G2/G3). - geocode endpoints (stations/stops/terminals/rental desks) on confirm via Nominatim; mapper now emits coordless named endpoints and confirm persists only the geocoded ones (G6). - assign every dated booking to the nearest trip day so it still shows when slightly out of range, and keep hotel accommodation from vanishing when a check date misses (G5/G10). - fix bus mislabelled as train + add bus_number metadata (G7/G8), flag malformed boats (G9), accept root start/end time for events (G11). - raise the local-LLM timeout to 300s for CPU-only Ollama.
client: the import call inherited the global 8s axios timeout and aborted long LLM extractions even though the server finished it; remove the timeout. server: raise the OpenAI-compatible LLM timeout 60s->180s (a cold Ollama model can take ~45s to first token). server: cap extracted text to 8000 chars before the LLM - multi-page T&C tails (30k+ chars) overflowed the context window, truncating the relevant head and making CPU inference crawl; booking details sit at the top.
Booking import only fell back to the LLM when each user flipped an 'always retry with AI' toggle, so by default files kitinerary returned nothing for just failed. Run the fallback automatically whenever the AI Parsing addon is on (fallback-on-empty); drop the now-redundant per-user toggle and its setting.
The first-run seeder only applies ADMIN_EMAIL/ADMIN_PASSWORD on an empty
database and then silently ignores them. People add the vars after the first
boot, or pull a fresh image without clearing ./data, restart, and cannot log
in with no hint why (#1339). The default is a generated password (not the
.env.example placeholder), printed once in the first-run box. Now: warn loudly
when the vars are set but a user already exists, and warn on a partial
(one-of-two) config instead of quietly falling back.
Also ship the reset-admin recovery script in the image -- it was never COPYed in
despite the wiki referencing it. node server/reset-admin.js resets/creates
admin@trek.local with a generated password (RESET_ADMIN_EMAIL/RESET_ADMIN_PASSWORD
overridable), picks a free username so it cannot trip UNIQUE(username), and sets
must_change_password.
The 3.1.3 docker build failed on linux/amd64 while bundling the client
(arm64 happened to pass): vite 8.0.16 pins rolldown 1.0.3, but tsdown
pulls rolldown 1.1.2, and on amd64 npm hoists the 1.1.2 native binding so
vite's 1.0.3 rolldown loaded a mismatched one ("builtin:vite-wasm-fallback
does not match any variant of BindingBuiltinPluginName"). Moving the client
to vite 8.1.0, which expects rolldown 1.1.2, lines the bundler up with the
hoisted binding. Verified by building the client in a clean linux/amd64
node:22 container.
A member of one trip could point a file at a reservation, place or
day-assignment belonging to another, private trip — on upload, on a
metadata update, or through the file-link endpoint. The reservation join
in the file list and the links list then returned that trip's reservation
title, disclosing it across the trip boundary and letting an attacker
enumerate foreign reservation titles by their id.
The file already had to belong to the caller's trip; now the linked
reservation/place/assignment must too. findForeignLinkTarget checks each
supplied id against the trip (assignments via day -> trip) and the upload,
update and link handlers reject a cross-trip reference with 400 before it
is stored. Same-trip links and clearing a link are unchanged.
The 3.0 "what's new" notices have served their purpose, so swap them for a single thank-you notice that comes back once on every fresh install and version bump. It carries Buy Me a Coffee and Ko-fi buttons and only shows on desktop. Adds a per-version recurring mode (new dismissed_app_version column) plus external-link CTAs to support it; the 3.0.14 whitespace-collision admin notice stays active.
Settle-up transfers are stored as fixed amounts, but a foreign-currency expense was re-converted with live rates on every settlement calc. When the rate drifted, the fixed transfer no longer cancelled the re-valued expense and a few-cent residual re-opened the settled position. Foreign-currency expenses now freeze the live rate at entry time into the existing budget_items.exchange_rate column, and the settlement converts with that frozen rate when working in the trip currency. Legacy rows (exchange_rate = 1) keep using live rates, so historical data is unchanged until re-edited; rate fetch failures fall back to live rates.
AirTrail returns each airline as {icao, iata, name}, but the import reduced it to the ICAO/IATA code, so an imported flight showed e.g. 'EWG' instead of 'Eurowings'. The picker and the stored reservation now use the airline name (falling back to the code when AirTrail has none). The raw code is kept in metadata.airline_code so the writeback to AirTrail still sends a code, not a name (#1240), and the change-detection snapshot hash stays on the code so existing flights don't spuriously re-sync.
Changing a trip's start date positionally re-dates the day rows (keeping their ids), so a dated booking's day_id stayed glued to a now-re-dated day and the booking visually shifted by the offset — until you re-opened and saved it. After a date-range change, non-hotel bookings are now re-anchored to the day matching their absolute reservation_time (the same derivation create/update already use). Bookings whose date falls outside the new range are left untouched; hotels and the relative positional shift of places/notes are unaffected.
On a GL map (Mapbox Standard) the basemap labels fell back to the browser/OS locale, so place and country names showed stacked in several scripts (e.g. 'India / भारत / India') regardless of the chosen language. Pin Mapbox Standard's basemap label language to the user's UI language via the basemap 'language' config property, mapping the few TREK codes that differ (br->pt, gr->el, zh/zhTw->zh-Hans/zh-Hant). Applies to both the trip map and the journey map; classic and MapLibre styles are left unchanged.
Pasting a single-place Google Maps share link (.../maps/place/...) into the list import failed with a cryptic 'Could not extract list ID from URL'. When the link is a single place it now returns a clear message telling the user to paste it into the place search box instead; other unrecognised URLs keep the existing list-link message.
getCountryFromCoords picked the country with the smallest bounding box containing the point, so a place just across a border (e.g. Strasbourg, which sits inside both the FR and DE boxes) landed in the wrong, smaller-box country. When more than one country box matches, it now disambiguates with the real admin0 polygon via point-in-polygon, smallest-box-first; a micro-territory with no admin0 polygon (HK, MO, SM, VA, ...) keeps the smallest-box win, and an unmatched point falls back to the old behaviour. The common single-candidate case is unchanged.
Asserts the route tools appear for one located place when a bookend accommodation exists, and stay hidden without one, guarding the #1330 visibility change.
Renders the public share page in German with a titleless day and asserts the i18n label 'Tag 1', guarding the t('dayplan.dayN') fix against a regression to a hardcoded English string.
The currency and timezone widgets stored their state only in browser localStorage, so a (docker) upgrade that clears site storage reset them to defaults — unlike every other preference, which is saved server-side. Persist them through the per-user settings store (no schema change; the settings table takes arbitrary keys) and migrate any existing localStorage values on first load so users keep their picks. dashboard_timezones is left unset by default so the widget can tell 'never chosen' from an explicitly emptied list.
Maurice