DEL-33165: Add Melia 1, split Models and Languages pages#258
DEL-33165: Add Melia 1, split Models and Languages pages#258cabbage-ice-cream wants to merge 9 commits into
Conversation
- Add new Models page at /speech-to-text/models covering Enhanced, Standard, and Melia 1 (DEL-33165) - Replace Languages and models page with Languages (slug unchanged) - Add Melia 1 multilingual subsections to Batch Input and Output pages - Add Models nav item above Languages (DEL-33183) - Update inbound links to moved/renamed anchors (DEL-33182) - Add melia and Arabic example word to custom-words.txt Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
|
||
| <SchemaNode schema={batchSchema.definitions.JobConfig} /> | ||
|
|
||
| ### Multilingual transcription with Melia 1 |
There was a problem hiding this comment.
This is very hard to find - should be visible in the table of contents imo.
Should also be linked to via the feature comparison table on the models page.
There was a problem hiding this comment.
Done — raised the page's toc_max_heading_level to 3 so the section shows in the TOC, and linked it from the comparison table's "Automatic multilingual" cell on the Models page.
| | Turnaround | Fast | Fastest | Fast | | ||
| | Processing modes | Batch and Realtime | Batch and Realtime | Batch | | ||
| | Regions | EU, US, AUS | EU, US, AUS | EU, US | | ||
| | Language handling | Selected language or pack | Selected language or pack | Automatic multilingual | |
There was a problem hiding this comment.
Should link to batch lang ID and language hints as appropriate
There was a problem hiding this comment.
Done — "Selected language or pack" now links to Languages with an auto-detect link to batch Language Identification, and "Automatic multilingual" links to the language hints section on the Input page.
|
|
||
| Melia 1 is available for Batch transcription in the EU and US regions only. It is not available in the Australia (AU1) region. | ||
|
|
||
| | Region | Endpoint | |
There was a problem hiding this comment.
Should this instead be on the authentication page? Or at lease linked to from there?
There was a problem hiding this comment.
Kept the table here so the Melia 1 section stays self-contained, and added a line under Supported endpoints on the authentication page noting Melia 1's regional availability with a link back here.
| | Processing modes | Batch and Realtime | Batch and Realtime | Batch | | ||
| | Regions | EU, US, AUS | EU, US, AUS | EU, US | | ||
| | Language handling | Selected language or pack | Selected language or pack | Automatic multilingual | | ||
| | Custom dictionary | ✅ | ✅ | Not yet | |
There was a problem hiding this comment.
Do we need notes / warning on the relevant pages to say that these aren't yet supported on Melia?
There was a problem hiding this comment.
Agreed in principle — but that's ~10 feature pages of new content, and Melia 1's feature support is expanding, so those notes need a maintenance plan. Proposing we raise a follow-up ticket for it rather than grow this PR. The comparison table on the Models page is the canonical list in the meantime.
There was a problem hiding this comment.
Raised https://speechmatics.atlassian.net/browse/DEL-33243 to track this.
|
|
||
| Melia 1 matches the Enhanced and Standard models for core transcription features, including diarization, word timings, punctuation, notifications, and output locale. It does not yet support the following features, which are available with the Enhanced and Standard models: | ||
|
|
||
| - Custom vocabulary and formatting: custom dictionary, find and replace, spoken form output, profanity tagging |
There was a problem hiding this comment.
I feel like this section should be part of the compare the models section, otherwise users have to check multiple places for this information
There was a problem hiding this comment.
Done — moved the unsupported-features list (and the early-access note) up into Compare the models, so the table and the detail are in one place. The Melia 1 section now links back to it.
| ``` | ||
| To choose a transcription model, refer to [Models](/speech-to-text/models). | ||
|
|
||
| The languages, packs, and options on this page apply to the Enhanced and Standard models. The Melia 1 model is multilingual: it transcribes the individual languages listed here and switches between them automatically, without language selection. It does not support the `auto` option, the bilingual and multi-language packs, or translation. For Melia 1, refer to [Models](/speech-to-text/models). |
There was a problem hiding this comment.
These language codes are also relevant for melia - it shows which languages can be transcribed by the model, as well as the codes to use for language hints.
Note that the bilingual codes arent relevant for melia.
There was a problem hiding this comment.
Done — the intro now says the language codes double as language hints for Melia 1 (linked to the Input section), and clarifies it's the bilingual/multi-language pack codes specifically that don't apply.
| | Norwegian Bokmål (no) | Norwegian Nynorsk (nn) | | ||
|
|
||
| ## Multilingual speech-to-text | ||
| ## Bilingual and multi-language packs |
There was a problem hiding this comment.
Would be a good place to spotlight and direct users to Melia a bit more prominently IMO - its pretty hidden away currently.
There was a problem hiding this comment.
Done — promoted that pointer to a note at the top of the packs section, same wording, much more visible.
- Show H3 sections in the Input page TOC (toc_max_heading_level 2 -> 3) - Link comparison table cells to Languages, Language Identification, and the new Input multilingual section - Deep-link the Input/Output references on the Models page to the new section anchors - Note Melia 1 regional availability under Supported endpoints on the authentication page, linking to Models Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…/speechmatics/docs into del-33165-melia-models-languages
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…1 on Languages page - Move the Melia 1 unsupported-features list into Compare the models so feature differences live in one place - Note that language codes double as Melia 1 language hints in the Languages page intro - Promote the Melia 1 pointer in the packs section to a note at the top of the section Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…ion not yet on Melia 1 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Summary
Implements the approved Melia 1 model docs content:
docs/speech-to-text/models.mdx) comparing Enhanced, Standard, and Melia 1, with model selection, Melia 1 details, and the operating points deprecation note (DEL-33165)Redirects and anchors (DEL-33182)
Fragment redirects are not implementable: Vercel redirects are path-based and URL fragments never reach the server, so neither
/speech-to-text/languages#operating-points→/speech-to-text/modelsnor an#multilingual-speech-to-text→#bilingual-and-multi-language-packsredirect can be added.vercel.jsonis unchanged (no page paths changed). External links to the old anchors will land at the top of the Languages page, which now points readers to Models in its intro.All internal inbound links were updated instead:
#multilingual-speech-to-text→#bilingual-and-multi-language-packs:features/feature-discovery.mdx(×2),deployments/index.md,deployments/container/cpu-speech-to-text.mdx#operating-points→/speech-to-text/models:features/audio-filtering.mdx(note: this anchor did not exist on the current Languages page either — the link was already stale)languages#models→/speech-to-text/models:deployments/container/gpu-speech-to-text.mdx,deployments/container/cpu-speech-to-text.mdx— the old### Modelsheading disappears with this change, so these would otherwise become broken anchorsAdaptations to repo conventions (no wording changes beyond these)
docs.speechmatics.comURLs converted to root-relative linkskeywordsfrontmatter on the Languages page preserved;title/descriptiontaken from the approved contentmeliaand the Arabic example word (مرحبا) added tocustom-words.txtfor cspellKnown open items — intentional, not oversights
language_pack_infoper-language object shape is an open question for the spec owner; the authored example stands as written.Verification
npm run buildpasses (broken internal links throw; none found). The 5 broken-anchor warnings in the build log are pre-existing onmainand unrelated to these pages.Quicklinkswarning on the Output page also exists onmain).Terminology note for a follow-up:
deployments/index.mdlabels the packs feature "Multilingual speech to text" andcpu-speech-to-text.mdxsays "multi-lingual transcription" — under the new terminology ("multilingual" reserved for Melia 1) these labels are now off, but renaming them was out of scope here. Only their link targets were updated.🤖 Generated with Claude Code