Adds the admin fixture the Search Sources screen needed, and repairs the reason
no browser-driving e2e test could log in at all.
The sign-in failure first. The suite drove the app over http on a container
hostname, which is not a secure context, so the browser provides no
crypto.randomUUID. AccountBoundary calls it to mint a session generation on
every sign-in; the call threw, the boot handler's catch swallowed it, and every
test landed on the login screen holding a perfectly valid session. Measured:
isSecureContext false and randomUUID undefined on
http://pediatric-ai-scribe-e2e:3000, both true on http://127.0.0.1:3553, where
boundary.enter() returns true and the app enters.
Chrome's --unsafely-treat-insecure-origin-as-secure was tried first and does not
work: Playwright rejects the --user-data-dir it must be paired with, and the
flag alone leaves isSecureContext false. Loopback needs no flags, so the runner
now uses the host network and the published port.
The seed is new. The e2e user was a registration someone did by hand once that
the shared Postgres happened to keep — enough to log in and no more. There was
no admin account, so nothing under /api/admin could be tested through a real
request, which is how the Search Sources card came to be verified by reading its
markup. e2e/seed.js creates both accounts and reconciles an existing one, so a
leftover with the wrong role cannot fail the suite for a reason unrelated to the
code. It resets passwords and grants admin, so it refuses any address outside
@ped-ai.test. The runner seeds before it tests.
The new spec covers what markup-reading could not: that an ordinary account is
refused the settings and never offered the Admin menu item, that no API key
comes back readable, that the Test button reports each source separately, and
that every control the save handler reads exists in a real render. Each account
gets its own browser context, because AccountBoundary allows one owner per
document and freezing the page on a second is the behaviour, not a bug.
10/10 pass on both projects. Two unit tests pin the loopback requirement and the
seed's domain guard so neither can be undone quietly.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
PubMed joins web search as an optional source for a generated resource: a
literature search on the topic, with abstracts, cited by PMID in References.
Off by default, admin-enabled, with its own optional API key (NCBI raises the
rate limit from 3/sec to 10/sec; it works without one).
Neither search is a tool any more, and that is the point. Offering them as
function calls meant the model decided whether to search, and with a prompt
ending "Output ONLY Pandoc markdown" it decided not to — every time, with and
without corpus grounding, no matter how the tool description was worded.
Calling callAI with the tool directly produced a correct pubmed_search call, so
the plumbing was never the problem. The search only ever needed the topic, and
the route knows the topic before it calls the model, so both searches now run up
front and their results go into the prompt as findings, exactly the way corpus
excerpts do. Ticking the box now means the search happened.
Verified live against deepseek-v4-flash: 30 corpus excerpts and 6 PubMed
results, and a References slide carrying both the library sources and four real
PMIDs (29562151, 38506440, 35721052, 28814254).
Three fixes to illustration, which had never once fired:
- The dispatch call had been lost in a refactor. The tool was still offered, the
model still called it, and the call was dropped, so no job was ever enqueued.
- imageContext was passed as a bare topic string where dispatch expects
{ request, history }, which made the bound request undefined.
- The prompt never mentioned the tool existed while explicitly demanding only
markdown — the same suppression that killed the searches. It now says an
illustration is available and that calling it is not a violation of that rule.
my_resources is its own image workflow rather than a reuse of learning_hub,
because generated_image_links only accepts learning_hub assets, and that is
exactly the barrier that keeps a private illustration out of published content.
The illustration renders in the panel, rather than a toast pointing at an image
history this feature does not have.
Verified end to end: job queued, rendered, and the asset served to its owner as
a correctly labelled subglottic-anatomy teaching diagram.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
It asked for 8 while the reranker capped everything at 12, so 8 is what it ever
actually received — and its answers were judged good on that. Now that the cap
defers to the caller, 12 is the number worth asking for: it is what the corpus
has been tuned against, and the reranker still decides which 12.
Verified live: the assistant returns 12 sources.
Also records PubMed as a search source of its own rather than a provider option
under web search. It returns structured records — title, journal, year, PMID,
abstract — so a reference can be exact instead of reconstructed from a page
title, and a model should be able to reach for "the literature" distinctly from
"the web".
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The one feature here that sends text outside the building, so the defaults are
the careful ones: disabled unless an administrator turns it on, opt-in per
generation even then, and the option is hidden entirely rather than shown as
something a user can tick and be refused.
Only the search query leaves. Library excerpts, the generated resource and
anything about the user never do. Both screens say so plainly, because a topic
typed while drafting clinical material can carry clinical detail and the
provider keeps its own logs.
Four providers behind one shape, so swapping changes nothing downstream: Tavily,
Serper over Google, Brave, and SearXNG — the only one where the query does not
reach a commercial third party at all, which is why it is worth supporting even
though it needs somewhere to run.
The tool description says when NOT to search, because a model handed a search
tool will reach for it constantly: not for settled clinical knowledge, which is
what the indexed library is for, and one search per resource. That last one is
enforced in the route with toolChoice: 'none' on the continuation rather than
trusted to the model.
A failed search never fails a generation — same contract as corpus retrieval.
The resource is written without it and the response says what was searched for
and what came back, so a query that left the network is visible rather than
silent.
The API key is masked on read and preserved when the field is left blank, the
handling the OIDC client secret already gets, so changing provider cannot
silently wipe a working key.
Verified on the running instance: with nothing configured, webSearchAvailable is
false, and a request asking for it anyway is ignored rather than honoured.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
A generated deck was losing content. pandoc writes a bare <a:bodyPr/> on every
shape, which leaves the body with no autofit even though the slide master has
one, so a slide with too much on it is cut off mid-sentence and the rest is not
rendered at all. Reproduced and counted: eight bullets went in, three came out,
the third ending mid-word.
Every generated deck now carries <a:normAutofit/> on its body placeholders. No
fontScale, deliberately — the renderer works out the reduction, so a slide that
already fits is untouched, where a fixed scale would shrink all of them. The
same eight bullets now fit with nothing in the bottom 6% of the slide.
This is a floor, not a licence to overcrowd. The prompt still asks for one idea
per slide; this stops a long one becoming unreadable.
Also: an article is no longer offered as PowerPoint. A deck of paragraphs is not
a presentation. Word and PDF suit either kind, and the route refuses the
combination rather than relying on the button being absent.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Answering a fair question: needing two settings in two repositories to agree
before one number changes is not a design anyone chose. It is two codebases each
assuming it owned the decision, and the symptom was a caller asking for 30
excerpts and silently receiving 12.
rerank_results computes min(reranker_top_k or limit, limit), so RERANKER_TOP_K=0
reads as "however many the caller asked for". The app setting is now the only
knob. Verified: Learning asks 30 and receives 30, the assistant asks 8 and
receives 8.
Zero costs nothing extra — the reranker is billed on documents sent, which is
candidate_limit and unchanged; top_n only decides how many come back. A real
number there is now what it should always have been: an optional hard ceiling
for when someone deliberately wants one, not an invisible default.
docs/retrieval-tuning.md covers the per-feature budgets, why the assistant's are
so much smaller than Learning's, that My Resources deliberately shares the
Learning budget, how to read what actually happened from the MCP logs and the
grounding field, and why raising these is not free.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The pathway existed but was reachable only by API. It now has a tab of its own
next to the Learning Hub — related, not the same thing, and sitting together is
how someone discovers the difference — visible to every signed-in user with no
role gate in the markup.
Generate a deck or an article, see everything you have made, download each as
PowerPoint, Word or PDF, delete what you no longer want. The screen says
"Private to you" and "Nobody else sees these", because the distinction from
published Learning content is the thing a person needs to understand before
typing a patient's condition into it.
Downloads are fetched rather than linked: an <a href> cannot carry the
Authorization header. The blob is saved under the filename the server chose and
the object URL is revoked afterwards. Resource titles come from a model, so rows
are built as elements and a title is only ever assigned to textContent.
The e2e stack now joins danvics_convert too. It could previously reach only
Postgres and Redis, so a PDF download failed there in a way production would
not — which did at least prove the degradation path works: with Gotenberg
unreachable the response is "PDF conversion is unavailable right now. PowerPoint
and Word still work", and the other two formats download unaffected.
Verified in a browser as an ordinary user: the tab appears and opens, the form
swaps slide count for word count when the format changes, the library lists
their own work, and pptx, docx and pdf all download with sensible filenames
(36360, 13285 and 68310 bytes).
Also documents retrieval sizing in docs/retrieval-tuning.md — the per-feature
budgets, and RERANKER_TOP_K, which caps all of them and had until now appeared
in no configuration file at all.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Learning is moderator-owned: content published into categories that everyone
sees, behind router.use(moderatorMiddleware). That is right for institutional
material and wrong as the only way in — an ordinary user could not generate
anything at all.
So this is a separate pathway rather than a loosening of that one. Learning is
untouched; the moderator gate stays exactly where it was. A signed-in user can
generate a deck or an article for their own use, keep it, refine it and export
it, and nobody else ever sees it.
Private by construction. Every statement filters on the owner and there is no
route that returns another person's work, which a test asserts statement by
statement rather than trusting. The foreign key cascades, so deleting an account
takes its drafts with it. There is no category, no publish state and no sharing:
adding sharing later should be a deliberate feature, not something that leaks
out of a forgotten WHERE clause.
Markdown is the artifact. Every format is rendered from it on demand — pptx and
docx by pandoc, both carrying the house reference deck, and PDF by Gotenberg,
whose LibreOffice preserves a deck's layout in a way rendering from markdown
would not. That is what makes "add a slide on when to admit" a text edit rather
than a binary patch.
Gotenberg was published on the host but on a network of its own, so reaching it
from a container went out and back through the host gateway. It now joins
danvics_convert, owned by danvics-net like the others. PDF is the one export
allowed to fail: if that service is down, the deck and the document still
download and the error says which.
Verified end to end as a plain user: the moderator route still refuses with 403,
generation returned a deck grounded on 12 corpus excerpts, the library lists
only their own, pptx/docx/pdf all downloaded valid, "add a slide on when to
admit" put the slide in the right place and left References last, and an
unauthenticated request gets 401 while someone else's id gets 404.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The route accepted useCorpus and reported grounding, but nothing in the admin
screen sent the flag or showed the result — so the feature existed and was
unreachable.
Opt-out in the UI rather than opt-in. For clinical teaching the library is
nearly always the right source, so someone who never notices the checkbox
should get the grounded version. The help text explains when to turn it OFF,
which is the non-obvious case: a topic the library does not cover is better
written without grounding than padded with the nearest unrelated excerpts.
Afterwards it says what happened — "Written from 12 library excerpts", or "Not
grounded — nothing indexed matched. Written from the model alone." Ungrounded
material presented as grounded is the failure worth preventing here, so the
wording never implies the library was used when it was not.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
A slide carrying [1] markers is unreadable from the back of a room, and an
article that cites inline reads as a paper rather than as teaching material. The
model is now told explicitly not to cite in the body — no bracketed numbers, no
parenthetical "(Nelson, p. 2604)" inside sentences — and to put everything it
drew on in a References section at the end, which in a presentation is the final
slide.
Checked rather than assumed: a six-slide deck generated through the grounded
path contains zero in-text citation markers, and ends with a References slide.
The prose keeps the specificity that grounding is for — bilirubin produced at
two to three times the adult rate, conjugation immature until about two weeks,
thresholds in mg/dL — without a single marker interrupting it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The grounding metadata went to the API response and the logs, which is no use to
someone holding the deck. A teaching resource shown to trainees should carry its
own provenance, so a grounded one now ends with a References section — the final
slide in a presentation — listing the library excerpts it actually used, by
title and page.
Restricted deliberately: only excerpts actually drawn on, nothing invented. That
was worth checking rather than trusting. Generated a deck and compared every
citation against the source metadata: "Kliegman R. Nelson Textbook of
Pediatrics, 22nd ed., 2024, p. 2604" against a stored title of "Kliegman R.
Nelson Textbook of Pediatrics 2-Volume Set 22ed 2024" at page 2604, and the same
for Fleisher & Ludwig, Rosen's, Understanding Pathophysiology and the AAP
compendium. The model reformatted filename-derived titles into readable
citations using only what it was given.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Learning generated everything from the model alone. A deck on bronchiolitis was
whatever the model remembered about bronchiolitis, with no connection to the
documents this institution actually indexed — while the assistant had been
searching that corpus all along.
Same collection, deliberately. mcp_bge_m3_1024 is already embedded with
openrouter-bge-m3 at 1024 dimensions; a second index over the same documents
with the same embedder would be a copy that drifts. What differs is the budget:
a chat answer wants a few tight excerpts because the reader is waiting, a
teaching resource synthesises a whole topic. So learning.search_limit and
learning.context_chars default to 30 and 2500 against the assistant's 8 and
1400, and are separate keys so tuning one cannot move the other.
Not unbounded, though. "No limit" only moves the ceiling from a setting to the
model's context window, where overflow truncates the middle of the prompt
silently — the worst place to lose source material. 60 results and 8000
characters per excerpt are the caps.
Opt in per generation: a resource on something the library does not cover is
better written without it than padded with the nearest unrelated excerpts.
Retrieval never fails a generation — the resource is then written from the model
alone, which is what happened before this existed — and every response reports
what it was grounded on, so a caller can say "24 excerpts" or "the library had
nothing on this" rather than quietly serving ungrounded material.
Verified against the live corpus: bronchiolitis, neonatal jaundice and febrile
seizure each returned 12 excerpts and ~23k characters from Nelson, Rudolph and
the Pediatric Clinical Practice Guidelines. A deck generated through the full
chain came back with textbook specificity that is not general recall —
bronchiolar diameter, birth-weight thresholds, the full pathogen list.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
npm audit reports 0 vulnerabilities. It reported 2 high and 2 moderate this
morning.
@google-cloud/vertexai was the last source of findings — gaxios and a uuid with
a missing buffer bounds check, neither reachable in this deployment because
GOOGLE_VERTEX_PROJECT is unset and the require sits inside that check. Dormant
is not the same as gone, and the provider is available through the gateway
anyway, so the direct path has been removed rather than left to rot:
- the SDK client and callVertex, which without the package could never run
- the dispatch and discovery branches that reached them
- VERTEX_MODELS, a list of ids nothing could route any more, and the two
places in adminConfig that concatenated it into the built-in set
- the health endpoint's vertex line, and the env vars documented for it
AI_PROVIDER=vertex now says where to configure the model instead of quietly
becoming something else. The Google STT and TTS paths keyed off the same
variable are untouched; neither ever used this SDK.
Verified after deploy: provider litellm, the assistant answers with 8 sources,
/api/models returns 10, and @aws-sdk/s3-request-presigner — which documents.js
needs for presigned MinIO URLs — is still declared and resolvable.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Generated a deck with ds-deepseek-v4-flash and rendered it to look at. The model
produced exactly the six headings it was asked for; the deck came out with eight
slides. The extra ones were pandoc's, not the model's.
Two rules, both found by rendering rather than reading:
- pandoc splits a slide after a table. Anything following one becomes a new
slide with no title — that was the stray "Key differentials to consider:"
slide floating with no heading.
- A table with no blank line before it is not parsed as a table at all. It
renders as literal pipe characters in the preceding paragraph.
And one that was visible on the slide itself: a nested ordered list inside a
bullet ran off the bottom.
None of these are the model failing. A cheap model writes perfectly good slide
markdown — bold, italics, nested lists and a table with a subscript all came
through correctly. It just needs to be told the shape the renderer wants.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
pptxgenjs is gone, and with it 269 lines of hand-rolled markdown parsing.
It stretched every image. Reading the slide XML it emitted shows why: it writes
the target box verbatim with <a:stretch/> and a no-op srcRect, so a 200x800
image handed an 11.8x3.9 box came out 1:4 squashed to 3:1. It could not do
better — it never measures an image, and its own getSizeFromImage is commented
out and marked "currently unused", reaching for a package called sizeof that
does not exist. pandoc measures them: a 300x175 source renders at aspect 1.714
and a 160x360 at 0.445, verified by rendering the deck to PDF and looking at it.
Tables, ordered and unordered lists, bold, italic and subscripts all come out
natively, and the fonts, palette and slide layouts come from
assets/learning/slides-reference.pptx. Design now lives in that file: restyling
the decks means editing it in PowerPoint, not editing this route.
Only images the requester owns can reach a deck. pandoc resolves an image link
against the filesystem, so a markdown link naming any local path would read that
file into the presentation. Images are fetched by id through the ownership
check, written into a per-request temporary directory under names we choose, and
every image link that did not resolve is removed rather than passed through. The
directory is removed in a finally block, and the conversion has a 60s timeout so
it cannot hang a request.
pandoc is in the image rather than a sidecar, because an export must not fail
for reasons outside this container. It costs 197MB (307 -> 504).
Removing pptxgenjs also removed image-size, and with it both high-severity
advisories — GHSA-w3rx-r6r6-pgpr and GHSA-5p2g-fcmc-qvqq, ICNS/JXL/HEIF parser
denial of service, ranged <=2.0.2 with no fixed release to upgrade to. npm audit
goes from 2 high and 2 moderate to 2 moderate.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Every generated image in an exported deck was distorted. PowerPoint scales an
image to whatever extent it is given, and the export handed it the content box
verbatim. pptxgenjs has a `sizing: contain` option that looks like it solves
this; reading the emitted slide XML shows it does not — a 200x800 image in an
11.8x3.9 box came out as cx=10789920 cy=3566160 with <a:stretch/>, stretched
from 1:4 to 3:1. It cannot do better: it never measures the image, and its own
getSizeFromImage is commented out and marked "currently unused".
So the export measures the image itself and hands PowerPoint an extent that
already has the right shape, centred in the space available. Verified: a 200x800
image now places 0.97x3.90 and a 4x3 places 5.20x3.90, both matching their
source aspect exactly, neither overflowing. An image that cannot be measured
keeps the old behaviour rather than failing the export.
image-size becomes a real dependency rather than one borrowed transitively, and
an override collapses it to a single copy — pptxgenjs declares it but the string
appears in none of its four shipped bundles, so npm was placing a second copy in
the production image that nothing could load.
Its ICNS, JXL and HEIF parsers have open denial-of-service advisories against
every published version (GHSA-w3rx-r6r6-pgpr, GHSA-5p2g-fcmc-qvqq, both ranged
<=2.0.2), so there is no release to upgrade to. They are disabled instead:
this application measures PNG, JPEG, WebP and GIF and nothing else. An ICNS
buffer is now refused and falls back to the box rather than entering the parser.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Retrying a kept recording transcribed it and put the text on the clipboard,
leaving you to find the right tab and paste. The app already had the answer: the
module was recorded with the audio. Retry now opens that tab and puts the text
in its transcript box.
Two things had to be true first.
The module was not actually being recorded. transcribeAudio never sent one, so
the server stored its default for every upload — all 28 rows in audio_backups
said "recording", and a retry had nowhere to send anything back to. Each
module's call now tags its own upload.
And the names disagreed. The recorders tagged 'encounter', 'soap', 'dictation'
while the recording-started events said 'enc', 'sick', 'dict'. One table now
holds the mapping and resolves the aliases, so the recorder that tags the
upload, the backup row that labels it and the retry that delivers it cannot
drift apart again.
Existing text is appended to, never replaced: a retry usually recovers
something on top of a live transcript, and overwriting would lose the words the
browser did hear. The box only exists once its tab's markup has been fetched, so
delivery polls briefly rather than guessing a delay, and falls back to the
clipboard if the tab never opens. An empty result says so rather than claiming
success. Backup rows now name their source and the button reads "Retry into
SOAP Note" instead of "Retry".
Verified in a browser: enc resolves to encounter, delivery switched tabs and
produced 'existing live transcript\n\nRECOVERED TEXT'.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
/metrics answered on every public hostname — app.pedshub.com, peds.danvics.com
and scribe.pedshub.com — with 201 lines naming routes, traffic volumes,
event-loop timings and process internals. No credential in it, but a free map of
the application for anyone probing. Prometheus scrapes the container directly
over the Docker network and never goes through the proxy, so a request carrying
X-Forwarded-For is by definition not Prometheus and now gets a 404.
METRICS_TOKEN allows an explicit override; unset, it can never match.
Verified: 404 on all three hostnames, and up{job="ped-ai"} still 1.
The workspace launcher was blank. [hidden]{display:none!important}, added to
stop "Retry loading settings" showing permanently, outranked
`body.assistant-mode-workspace .assistant-workspace-view { display:block }` —
and that view is the one element toggled by CSS rather than by JavaScript
removing the attribute, so it could never be revealed again. The attribute is
gone from the markup; the class already defaults it to display:none, so it still
starts hidden. Everything else using the attribute is JS-toggled and unaffected.
Verified with a real login: 17 cards at 208x40 in an 860x240 grid.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Removed at the owner's request, entirely rather than switched off: the route's
allow-list, anonymous identity and flag lookup; the client's entry path, the
authFetch exception that let four endpoints out without an account, and the
workspace guard; the CSS, the in-page note, the admin flag and its save/load;
the test file and the assertions elsewhere that pinned it. Both settings rows
are deleted from app_settings.
Two things were checked rather than assumed on the way out. Removing the
anonymous identity collapsed every `if (!req.user.preview)` branch to its
authenticated side, so image tools, audit logging and citation storage now run
unconditionally — which is what they did before preview existed. And the route's
gate went back to a bare router.use(authMiddleware), which on a /api mount gates
every path below it in server.js; it is scoped to /clinical-assistant again, the
guard test catches it either way.
Verified after deploy: signed out, status, examples and chat all refuse with
401; signed in, chat still answers with 8 sources; extensions, encounters,
documents and admin remain shut.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
/assistant in the address bar, the preview class on the body, and a blank
screen. The app container starts hidden and enterApp is what reveals it;
enterPreview declined the login overlay without doing the same, which left a
hidden app under a hidden overlay. One line.
Verified in a browser: composer, chat view, example prompts and the preview note
are all visible and sized, with no page or console errors.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The server side has worked since this morning, but no browser could reach it.
authFetch rejects every /api request that has no account before it is sent,
and the four preview endpoints were not on its short list — so the status call
that decides whether to show the login screen never left the browser, and the
screen was always shown. The list now mirrors the server's own allow-list
exactly: status, examples, chat, chat/stream, and nothing else.
Preview now begins from any path. A visitor landing on the root met the login
wall while /assistant did not, which read as "preview doesn't work"; both now
enter the assistant, and the URL follows.
Reaching for anything that needs an account raises the sign-in screen through
one hook in authFetch rather than a check on every control — but only for
something the visitor did. The page also fetches saved chats and config in the
background on load, and the first version raised the screen for those too,
burying the assistant before a word was typed. The hook is gated on
navigator.userActivation.
The HIPAA notice is hidden on that screen in preview: it is an invitation to
sign in, not the compliance notice a clinician sees on first login.
Verified in a browser: landing on / and on /assistant both show the assistant
with no login wall and no HIPAA text; clicking Workspace raises sign-in.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
mcp-server_mcp-data carried the Compose project prefix, which tied its name to a
directory that has since been renamed to clinical-assist-deploy. It is now
clinical-assist-data, named explicitly in both stacks so it no longer depends on
a project name at all.
Migrated rather than renamed, because Docker cannot rename a volume: writers
stopped, 2.9GB copied, then verified before switching — 13265 files both sides,
3080478708 bytes of file content both sides, and an md5 over every file in
sorted order matching exactly. The old volume is kept, and is now referenced by
nothing, so it can be removed once this has run for a while.
Also removed mcp-server_qdrant-data: 1.7GB belonging to a Qdrant deployment that
no longer exists and which no container mounted.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The card holds more than one Save button and nothing said so. "Save image
settings" is injected directly above "Save model & retrieval settings", with no
indication of where one stops and the other starts, and the page saves nothing
automatically. It now states that up front, and the bottom button says which
settings it applies.
"Retry loading settings" sat beside Save looking like an ordinary control,
because it did: a bare button with the hidden attribute, which the browser's own
[hidden] rule could not hide once .btn-sm set a display. It is now inside an
error message that exists only on failure, says what failed, and says that
nothing typed has been lost.
The status line used to read "Settings ready." forever, which answers a question
nobody asks. It now reports the thing an admin actually wants to know when they
come back: whether the last save went through, and at what time. A toast is gone
in three seconds; this stays on the page.
The signed-out preview moves to Feature Flags, where it belongs. It was a second
checkbox under a row labelled "Sources", followed by two paragraphs, the first
about preview and the second about citations — so neither paragraph clearly
belonged to either checkbox. It is stored as feature.assistant_preview now, with
the old clinical_assistant.preview_enabled still honoured when the new key has
never been written. That also means an ordinary admin can toggle it under
ADMIN_LOCKDOWN: clinical_assistant.* is locked, and putting a day-to-day switch
behind host access was never the intent.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
"mcp" said nothing — this host runs several MCP servers. The clinical corpus
query service now also answers to clinical-mcp, and ped-ai asks for it by that
name. The bare service name stays an alias automatically, so anything still
pointing at http://mcp:8000 keeps working. Hyphens rather than underscores:
underscores are legal in Docker's embedded DNS but not in hostnames generally.
LITELLM_API_BASE goes back to being unset here, so OpenBao's
https://llm.danvics.com applies. The public hostname is the deliberate choice for
consistency across the estate. It is not a speed question — the gateway hop is
~19ms on calls taking hundreds of milliseconds, and describing that as "4x" was
a misleading way to put it. The real cost is that AI calls now depend on Caddy,
public DNS and edge TLS; one line here pins it back to the container network.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Both came out of using the app: the preview works but is all-or-nothing, and
signing in should be possible without a password without replacing it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
ped-ai joined open-webui_openwebui-network to resolve exactly one hostname:
libretranslate. LibreTranslate is its own Compose project but had no network of
its own, so it borrowed Open WebUI's — which meant a clinical app's startup was
coupled to an unrelated stack's lifecycle for a translation call.
LibreTranslate now owns danvics_translate and ped-ai joins that instead. It also
keeps its Open WebUI attachment, because Open WebUI is configured to use it
(webui.db refers to it) and cutting that would have broken translation there
silently. Verified after the move that Open WebUI still reaches it.
LITELLM_API_BASE is pinned to the container network again, now deliberately and
with the reason recorded: clinical traffic should not leave the host or depend on
Caddy, public DNS and edge TLS. Speed is not the reason — it is ~19ms on calls
taking hundreds of ms, and quoting that as "4x" was misleading. OpenBao still
stores the public URL for this key and Compose wins, so the vault value is dead
until someone changes it by hand; the app's AppRole is read-only by design.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
"Retry loading settings" was permanently on screen even though the settings had
loaded fine. The browser's [hidden]{display:none} lives in the user agent
stylesheet, so any author rule setting display beats it — .btn-sm sets
display:inline-flex, so `<button hidden class="btn-sm btn-ghost">` never hid.
Earlier fixes bolted style="display:none" onto individual elements one at a
time; one [hidden]{display:none!important} covers all fourteen.
LITELLM_API_BASE is no longer pinned in Compose. Compose env wins over OpenBao,
so hardcoding http://litellm:4000 silently overrode the vault's
https://llm.danvics.com and the intended configuration had never taken effect.
The entrypoint now applies 28 secrets rather than 27, with 2 deliberate
overrides rather than 3. Transcription, text-to-speech and the signed-out
preview were all checked through the public gateway afterwards.
Worth recording since it was the reason for the change: this clears no network
dependency. ped-ai still needs danvics_mcp for mcp, the assets network for
MinIO, the openwebui network for LibreTranslate and danvics_monitoring for Loki;
none of those has a public route. It also adds about 20ms per call.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Eleven routers are mounted on '/api' and called router.use(authMiddleware) with
no path. Mounted that way, the gate applies to every /api request that reaches
the router — including routes belonging to routers mounted further down
server.js. extensions.js did it from line 295; the assistant is mounted at 305.
So a signed-out request to /api/clinical-assistant/status was refused ten lines
before the preview middleware could look at it, whatever the admin setting said.
server.js line 250 already warned about this shape.
Each gate now names its own prefix, so a router protects its own routes and
nothing else. Verified afterwards that every namespace which must stay shut
still answers 401 signed out: extensions, encounters, memories, notes, diagrams,
generated images, image jobs, documents, audio backups, ED encounters,
don't-miss, patient education, billing, well visit, admin, transcribe and the
rest. Two of these routers were gating routes nobody realised they were gating.
Second defect in the same path: authMiddleware only ever looks for a token, so
calling it unconditionally after the preview identity had been assigned rejected
exactly the requests preview exists to serve. Only that identity may skip it;
authMiddleware stays strict everywhere else.
Preview now answers with a real cited answer, and stays as narrow as it was
designed to be — four allow-listed paths, no identity, nothing ownable. A test
now walks every /api router and fails on a blanket gate, which is how the last
six were found.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
"AI-Powered Clinical Documentation" named about a third of what is here.
Documentation is one pillar; the assistant and the learning hub are the others,
and none of it said pediatric, which is both true and the differentiator.
danvics_speech is gone from ped-ai. Everything ped-ai dials was checked against
every network it joins: assets, libretranslate, litellm, mcp and monitoring-loki
are the only internal hosts, and litellm answers to that alias on danvics_mcp
already. supertonic/kokoro/sherpa appear here only as LiteLLM model ids and
voice names, never as hostnames — speech goes through the gateway, exactly as
intended. Joining danvics_speech therefore bought no reachability and cost a
cold-boot dependency on local-speech-gateway, which had to have been up once
before ped-ai could start at all.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The recordings were never the problem. Six stored recordings were pulled back
out of object storage and examined: valid webm/opus, 3-5 seconds, 1.5-2s of
continuous speech-shaped audio each. Every one came back from
mistral-voxtral-mini-transcribe as an empty string, while the same model
transcribed synthesised speech perfectly — including a one-word clip, and
including that speech attenuated to the same level, so neither length nor
loudness explains it. Re-encoding to wav, mp3, flac, ogg and a remuxed webm
changed nothing; groq-whisper-large-v3-turbo transcribed all six. stt.model is
set to that now, and the real recording round-trips through /api/transcribe as
"Hello." instead of "".
So the server now says something when a model answers 200 with no words for a
non-trivial amount of audio. That silence is what made this look like lost
recordings; the log names the backup id, so the kept audio can be tried against
another model directly instead of suspecting the microphone.
Also: browsers report "audio/webm;codecs=opus", and deriving the extension by
splitting on "/" alone named the upload "audio.webm;codecs=opus". This gateway
tolerates it. A provider dispatching on extension would not.
And registration is one card again: enable it, decide whether it needs an
invitation, hand out codes — top to bottom. The invite-only switch sat in a
separate card far below the enable/disable toggle, which made one decision look
like two unrelated settings.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Reproducibility means two things here: the same commit builds the same image,
and the running container can be asked which commit it is.
- Base images are pinned by digest, not by tag. A tag moves; two builds of one
commit could otherwise differ. These are manifest-list digests, so buildx
still picks the right architecture.
- scripts/build-image.sh also writes ped-ai-local:<revision>, an immutable
name a deploy can refer to instead of chasing :latest. Its summary goes to
stderr so stdout stays the Compose invocation.
- Compose takes the image from PED_AI_IMAGE, so a deploy runs a specific
revision-tagged image while a local build still uses the local tag.
- scripts/deploy.sh pins that image in the file Compose interpolates from,
waits for health, then asks /api/build which revision is actually serving
and rolls back to the previous image if it does not match. Healthy is not
the same as running what you asked for. The rollback path was exercised.
- The entrypoint applies migrations before the app starts, so code and schema
arrive together. node-pg-migrate takes an advisory lock; losing it is not an
error, it waits and looks again, so a rolling restart does not fail. A real
migration failure stops the container rather than serving on a schema that
does not match the build. RUN_MIGRATIONS=false opts out.
- The Forgejo workflow builds through that same script, tags by full revision,
and has an opt-in deploy job. It refuses to run if the deploy directory has
uncommitted work rather than resetting over it.
The running image was labelled revision=unknown, and /api/build said "unknown",
because `docker compose up --build` never passes GIT_REVISION. That is exactly
the hole this closes.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Measured in a real browser against the app rather than reasoned about.
The transcript boxes are contenteditable divs, and an editable div zooms on
focus exactly like an <input>. The earlier 16px sweep covered input, textarea
and select, so every workspace tab still zoomed while the calculators did not —
which is exactly what was reported. Every focusable text control in every tab
now measures 16px at phone width; the count of ones below it is zero.
Three ways a recording could end with nothing to show for it:
- Safari supports none of the audio/webm types and throws NotSupportedError
when handed one. Six modules built their own recorder on resume with
"opus, else audio/webm", so resuming threw there and the recording stopped.
There is now one codec chain in the app, and no module constructs a
MediaRecorder of its own.
- audio-recorder-failed is dispatched on document, and the encounter tab
stopped its recording on any of them. The assistant's microphone failing
ended a consultation being recorded in another tab. The recorder now
travels with the event and the listener checks it is its own.
- The server answers {success:true, text:''} for silence, and five modules
assigned that straight into the transcript — emptying the box the browser
had been filling live. It reads as a recording that vanished. Text is now
required before overwriting, and a recording that captured nothing says so
instead of resetting the button over an empty box.
Also: the citation counters were registered on prom-client's default registry
while the app serves its own, so they were never scraped. They read zero at
/metrics now instead of being absent, which is what the Grafana panels need.
And the reference linter passes for the first time, so scripts/e2e.sh gets past
its preflight: KaTeX is vendored (it was referenced by the assistant's LaTeX
rendering but never shipped — three 404s a page load and no math), and the
JavaScript left behind by the removed image picker, saved-chats toggle, image
gallery and visual-output panel is gone.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Citation quality
- A citation naming a source that never came back is never rendered as a
link, so it appears as plain text and nobody learns it happened. It is now
measured on the server, where the answer and the sources both exist, so it
is seen whether or not a browser rendered it.
- Four Prometheus counters feed a Grafana dashboard (Ped-AI Citation
Quality): answers, citations written, answers affected, and individual
unresolved markers. Only answers with at least one unresolved citation are
stored, with the question and the titles retrieval returned, so an operator
can judge whether retrieval came back thin or the model over-cited. Rows
expire after 30 days: this is a quality signal, not a transcript log.
- Both answer paths are covered. /chat/stream is normal; /chat is the
fallback the client uses when streaming fails, so auditing only the first
would have hidden exactly the answers produced under failure.
- The tracker is resolved on demand and allowed to be absent. Seven test
files load this route with a hand-built list of permitted imports, and
adding a hard dependency would mean editing all seven — and the eighth
written later would break. Observation must never be able to fail an
answer, so a missing module simply means no tracking.
- Metric registration reuses an already-registered counter, because this
module can legitimately load twice in one process.
SSO settings on mobile
- Six rows were laid out inline: flex with a 160px label and an input that
would not shrink, so on a phone the row was wider than the screen with
nothing to scroll and no way to reach the rest. They use .admin-row now,
which already stacks below 640px. Verified at 390px and 360px: nothing
off-screen, no sideways overflow.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
With more than one admin, every setting in the panel was editable by all of
them — prompts, model policy, retrieval budgets, SMTP, email templates.
ADMIN_LOCKDOWN separates running the service from changing how it behaves.
It is an environment variable on purpose: a setting could be switched off by
the very admin it restrains, so lifting this needs host access and a restart.
The server is the control. One gate refuses configuration writes rather than
a check in each of the fifteen write routes, because that list grows and a
route added later would quietly miss it. Reads always pass — lockdown hides
nothing. Day-to-day operation stays available: invitations, announcements,
registration, feature flags, and the test endpoints, which persist nothing.
A setting invented later is locked until someone deliberately makes it
editable, rather than defaulting to open.
The panel disables what it cannot save and says why, but that is courtesy;
the refusal is what enforces it.
Two things this taught me, both fixed: my first version painted the panel
from an IIFE, which the module conventions forbid, and fetched the whole
config a second time just to read one flag — breaking the test that pins
admin loaders firing exactly once. The state now rides on the invites
response the panel already requests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
registration_enabled was a single switch: open to anyone, or closed to
everyone. This adds the setting an operator actually wants in between — open
to people you invited.
A code is single-use, expires (7 days by default, 90 maximum), and can be
revoked or deleted. It is stored hashed with only its last four characters
kept, because an invite grants account creation and a database dump should
not hand someone a working one. The code is readable exactly once, in the
response that creates it.
The claim is a single conditional UPDATE carrying every condition, so two
registrations racing the same code cannot both succeed. It happens after the
account exists, so a code is never spent on a failed registration — and if
the race is lost, the just-created account is removed rather than left behind
as a free registration. The rejection never says which of the four reasons
applied; distinguishing them would tell someone probing codes which guesses
were closer.
Codes avoid I, L, O and U so they survive being read aloud or copied off a
screen, and matching ignores case and separators.
The sign-up field appears only when the server says a code is required. The
admin card creates, lists, revokes and deletes, and carries the toggle.
Verified against the live database: create, claim, second claim refused,
unknown code refused, revoking a used code refused, delete. 684 tests pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
It documented 18 of 67, so anyone setting up a deployment had to find the
rest by reading source — including CORS_ORIGINS, REDIS_URL, NODE_ENV and the
whole Clinical Assistant retrieval group. Each entry now shows the default
the code falls back to, so an unset line reads as "this default" rather than
"broken".
Also documents the OpenBao switch: the entrypoint only fetches secrets when
OPENBAO_ADDR is set, a compose or env_file value always beats the OpenBao
copy, and turning it off means unsetting that one variable and putting the
keys here — with the command to list the key names without printing values.
A test now derives the set of variables from source, so this cannot drift
again.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
A marker names a source by the `number` dedupeSources assigns server-side.
Rendering looked it up as sources[n - 1], which works only while the array
order and the numbers agree. Nothing breaks that today, but it is an implicit
contract across a network boundary: any later filtering or reordering of the
list — hiding low-score sources, say — would point citations at the wrong
source silently, which is worse than not linking at all.
Matching on the number cannot drift. Positional lookup remains as the fallback
for a list whose entries carry no number.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
It was textContent, so formatting appeared literally. It now renders inline
Markdown only — bold, italic, code, links, breaks — because this text is
injected into every page and must not be able to introduce headings, images
or block layout that shifts the app around.
parseInline rather than parse, a tag allow-list, event-handler and style
attributes forbidden, data attributes off, and one innerHTML assignment that
can only be reached through DOMPurify. Missing either library, or any failure
while rendering, falls back to the literal text rather than to unsanitised
markup.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
CLINICAL_ASSISTANT.md described DeepL as an admin-configurable translation
provider with its own keys and a fallback between providers. There is no
DeepL client in the code: TRANSLATE_PROVIDERS is ['libretranslate'], and an
unrecognised setting silently falls back rather than being rejected — the
first wording I wrote claimed a rejection that does not happen either.
TODO.md now names clinical-storage-milvus and
personal-assistant-storage-milvus, and records the MinIO audio-backup work
as finished.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Written in the operator-handbook style of the Ops Docs collective (title,
last-updated line, short prose, tables, a quick-checks block, no secrets).
Ready to paste into the collective, or to publish once writing to Nextcloud
is permitted — that call was blocked by my permission layer.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
ped-ai-storage held the personal assistant index, and nextcloud-mcp-server
held ped-ai's clinical index and runs no MCP server at all — only etcd,
MinIO and Milvus. The names said the opposite of the truth:
ped-ai-storage -> personal-assistant-storage-milvus
nextcloud-mcp-server -> clinical-storage-milvus
Only the secret bind mounts here name those paths. Both compose files pin
`name:`, so containers, networks and volumes keep their existing names and
no data moves; that also means the container prefixes still read
ped-ai-storage_*, which can only change by copying each volume.
Verified: all three stacks' compose configs validate, ped-ai recreated with
the new secret paths and reads them, and the storage stack still resolves
its own containers from the renamed directory.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The nc_semantic_search alias is gone from the MCP server, so accepting it
here would point retrieval at a tool that no longer exists. A stale
override now stops the app at startup instead of silently retrieving
nothing.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
The app key carries a second policy covering only the audio-backups bucket,
so the already-mounted credentials serve both. Verified on the live stack:
storage=object, key recordings/<user>/<stamp>, round trip byte-identical,
and the object is removed with its row.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Three S3 configurations had grown separately — S3_* for documents,
GENERATED_IMAGES_S3_* for images, and AUDIO_BACKUPS_S3_* after them — with
different key names and their own client construction. That is why moving
storage meant hunting through several files.
src/utils/objectStorage.js now resolves settings for any purpose: its own
variables first, then the shared S3_* ones, with a per-purpose bucket name
(S3_BUCKET_AUDIO_BACKUPS). One endpoint plus three bucket names is enough
for the whole app, and a purpose that needs its own account still overrides
everything. Audio backups and documents use it; generated images keeps its
own tested storage module, whose variable names the resolver already
understands.
Nothing existing has to change: S3_ACCESS_KEY_ID, S3_SECRET_ACCESS_KEY and
the AWS_* fallbacks still resolve, and path-style addressing keeps each
purpose's previous default — off for documents, so a Backblaze endpoint
behaves as before, on where a custom endpoint implies MinIO. A _FILE
credential now always beats an inline one, so a mounted secret cannot be
shadowed by an inherited environment variable.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Same MinIO, its own bucket, as asked. The audio-backups bucket is created;
what remains is a MinIO IAM change, which needs root credentials — my
permission layer blocked that call, so it is a script to run rather than
something I applied.
The script is additive and reversible: it attaches a second policy covering
only the new bucket and carries the existing generated-images grant over
rather than replacing it (attaching only the new one would break image
storage). It prints the AUDIO_BACKUPS_S3_* values to set, and how to undo.
A bucket policy alone does not work here: MinIO evaluates the user policy
first and it denies by default. Verified — the app key gets AccessDenied on
the new bucket until its own policy allows it.
TODO records that, and the indexer being repointed from Documents to
Personal assistant so mail is finally reached.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Switching to the Assistant set window.location, which reloads the document
and silently ended any running recording. The assistant is a tab in the same
page and activateTab already rewrites the URL to /assistant, so the switch
now happens in place; the reload stays as a fallback. Verified in Chromium:
recorder still running, no reload, URL /assistant, assistant visible.
Signing out mid-recording used to end it with nothing kept. It now says so
first — "the audio will be saved for 24 hours so you can transcribe it
later" — and stores the audio either way, tagged with the module that
produced it ('encounter', 'soap', 'dictation'), which is what makes it
findable in Settings afterwards. Sign-out completes whether or not the save
worked, and the rescue never rejects, because the caller is on its way out.
Verified: the warning appears and the upload carries module=encounter.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
- docs/speech.md: a new Recording section (wake lock and why it is re-taken
on visibility, idempotent start, what happens when a recording ends by
itself, what signing out does), the retention rules, and step-by-step
instructions for moving audio backups to MinIO.
- Records that the model for a transcription is the user's choice first,
which is why the picker must only ever offer models the gateway has, and
that LITELLM_STT_MODELS is a fallback rather than a list known to work.
- api-reference: /api/transcribe takes `module` and returns `backupId`, and
the Audio Backups group is no longer failure-only.
- configuration.md and .env.example document AUDIO_BACKUPS_S3_*, preferring
the _FILE credential variants.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
Recording
- A screen wake lock is held for as long as a recording runs. Browsers drop
the lock whenever the page is hidden, so it is taken again on return —
without that, one glance away ended it for the session. The lock is
reference counted (two recorders cannot release each other's), never
requested while hidden (the request would just be rejected), and a denial
or an unsupported browser leaves the recording running.
- Signing out releases it and stops the recording; nothing is sent, because
the session that owned the audio is gone.
- start() on an already-running recorder is now a no-op instead of replacing
the MediaRecorder and silently dropping everything captured so far.
- A recording that ends by itself — recorder error, or the microphone taken
by another app, unplugged or revoked — takes the same path as pressing
Stop, so it is transcribed and stored rather than left in a tab that still
says "recording". Moving around the workspace already kept recording.
Retention
- Every recording is kept for 24 hours now, not only the ones whose
transcription failed. /api/transcribe already has the audio, so this costs
no second upload, and a storage failure is logged rather than thrown: it
must never lose the transcription someone is waiting for.
- One store (src/utils/audioBackupStore.js) is shared by /api/transcribe and
/api/audio-backups so the two cannot drift. Payload goes to object storage
when AUDIO_BACKUPS_S3_* is set and to the encrypted Postgres column
otherwise; metadata always stays in Postgres, so listing, ownership and
expiry behave the same either way. Object keys are scoped by owner, and
the expiry sweep deletes the object with the row.
Verified against the live database: round trip byte-identical, another user
reads null, 950 -> 48 bytes compressed, expired rows take their objects.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dv6sqaY6Vq3ChZHMem3cnU
