# 1c7/Translate-Subtitle-File (译幕 / Zimoo): BYOK Subtitle Translation for .srt, .ass and .vtt > A desktop and web tool that translates subtitle files using your own provider API key. The trade-off is explicit: cheapest possible per-character cost in exchange for setup work the README admits new users find complicated. **1c7/Translate-Subtitle-File** — 🔥（26年8月24号更新 v5.5.26 版）产品名："译幕"，用途：翻译字幕文件 .srt .ass .vtt。你只需要拖入字幕，点击翻译就行了。软件有网页版（zimoo.app）和桌面版。主要特点：用户可以自己填 API key。BYOK（Bring Your Own Key） - Repository: https://github.com/1c7/Translate-Subtitle-File - Website: https://zimoo.app - Stars: 2,634 · Forks: 195 - Language: Unknown - License: not declared - Published: 2026-09-28 · Updated: 2026-09-28 · Language: en - Canonical page: https://hysenlabs.com/projects/1c7-translate-subtitle-file ## The problem 译幕 solves, and the setup it charges for it Subtitle files are structured: each cue carries an index, a timecode and one or more lines of text. Machine-translating them with a general-purpose tool means either pasting text without timecodes and reassembling it by hand, or trusting a service that hides its pricing. 译幕 takes the file directly. The README states the supported formats are .srt, .ass and .vtt, and that the workflow is to drag a subtitle file in and press translate. The audience is narrower than the format list suggests. The README names two groups: individuals translating a tutorial or a video they want to understand roughly, and subtitle groups doing a rough machine pass before human editing. It also states machine translation accuracy sits somewhere in a 50% to 85% band and that anyone needing 信达雅 should hire a translator. That is an unusually candid scope statement, and it should be read as the product boundary rather than a disclaimer. The cost model is the reason the software exists. Instead of the project reselling translation, you configure a provider API key yourself. The README calls this BYOK and lists the benefit plainly: lowest price, no middleman cut, and you can use the provider's free quota, which it describes as typically one million characters per month. The stated downside is equally plain: you have to register and activate a service on the provider's site, and the README says first-time users will find this somewhat complicated. That is the whole trade in one sentence. ## How BYOK actually routes your subtitle text The architecture is a client that talks to translation provider APIs with credentials you supply. There is no project-operated translation backend in the described flow. When you start a job, the software sends the subtitle content to whichever provider you selected, and the provider bills your account directly. The README is explicit that the purchase value of the paid version is a graphical interface for reaching provider services directly, with no third party in between, and that the cost therefore cannot be lower. Because the software is a pass-through, its language coverage is inherited rather than implemented. The README lists English, Japanese, German, French, Korean, Russian, Spanish and Italian, then qualifies the list: if the translation provider supports a language, the software supports it. Do not read the list as a guarantee. Read it as a pointer to the provider's own documentation. Output is not fixed to bilingual subtitles. The README states you can output translation only by setting the format to 译文, or original plus translation by choosing 原文+译文. The commented-out sections of the README also describe per-provider character counting and per-provider quotas in the paid tier, which is the feature that matters if you are distributing work across several free quotas to stay inside them. The repository itself is thin at the top level: .gitignore, README.md and an app/ directory. The README references screenshots under app/image/, which is consistent with an Electron application, and the repository topics include electron and vuejs. There is no published source tree for the translation logic in what is visible here, so treat the mechanism as documented behaviour rather than something you can audit from the repository listing. ## Installing the desktop build and translating a first .srt file There is no package manager install documented, and the README contains no shell commands to copy. It points to the download page on the project site, and the desktop build is the thing you download from there. The web version at zimoo.app needs no install, but the README warns that it requires circumventing network restrictions to work normally, and recommends it anyway for the open-and-use convenience. Once the desktop application is running, the sequence the README describes is: configure a provider, drag in a subtitle file, choose the provider, choose source and target languages, then press start. Configuration is the step that gates everything else, and the README does not put it in the README itself. It links out to a separate documentation site: ```bash https://doc.tern.1c7.me/zh/ ``` That documentation is written in Chinese. If you do not read Chinese, budget for that before you start, because the provider registration walkthrough lives there and not in the repository. For the file you feed it, the README gives the accepted extensions directly: .srt, .ass and .vtt. The basic concepts section of the README spells out the full sequence, including that subtitle files carry the extensions .ass .vtt .srt, that you pick a provider from a dropdown, choose the source and target languages, and click the start button on the right to begin processing. After the job runs, the README says you get a translated subtitle file back, in bilingual form by default, or in translation-only form if you change the format setting to 译文. What you should check first is the timecode alignment on a few cues and the handling of any line that was already in the target language, because the README does not document either case. ## Where the BYOK design breaks down for a new user The honest limitation is stated by the project itself. The README's own pros-and-cons block lists the advantage as configuring your own API key for the cheapest price, and the disadvantage as needing configuration before the software is usable at all. There is no bundled trial key and no fallback provider. A user who installs the app and drags in a file without a provider account gets nothing. The second limitation is provider-side. Because the software is a pass-through, a provider outage, a quota exhaustion or a rate limit surfaces as your problem, not the project's. The README's own example of this pattern is character budgets: it notes that free quotas are typically one million characters per month, and that the paid tier's value includes per-provider character counting and per-provider caps so you do not exceed them. If you are on the free tier and you burn a quota, the README does not document what the application does next. The third is scope drift. The README's commented-out sections describe a speech-to-text feature with per-provider free quotas, and state it was disabled in version 5.3.0 because a single maintainer could not keep up with it, with the intention of adding features back gradually. Those sections are commented out, so treat speech-to-text as unavailable. If transcribing audio or video is your actual task, this is the wrong tool today. Finally, quality. The README puts machine translation accuracy between 50% and 85% and says so in the context of recommending human translation for high standards. For burned-in viewing of a technical talk that is often enough. For a subtitle group's final release it is not, and the README frames the intended use as rough processing before manual improvement. ## Alternatives, and how they differ in approach rather than in features The README lists several alternatives itself, and the differences are structural. 字幕酱 is a website rather than a desktop application, which the README notes approvingly because it needs no configuration. That is the direct opposite trade: you skip provider registration and pay a per-minute rate the README quotes as free for the first minute and from 0.2 yuan per minute thereafter, with the caveat that the figure was copied from the vendor site in August 2020 and current pricing should be checked. If setup complexity is why you are hesitating, this is the alternative that removes it. VideoSRT is Windows-only, which the README flags as a limitation against 译幕's macOS support. It supports Alibaba Cloud for speech recognition and Baidu and Tencent for translation, a narrower provider set than 译幕 claims. The README also concedes that VideoSRT's customer support is better, with two active 500-member QQ groups. On the command-line side, Autosub splits audio into chunks under 60 seconds and sends them through short-speech recognition APIs, then emits subtitle files. The README's objection is learning cost for non-programmers rather than capability. pyTranscriber wraps Autosub in a PyQt interface, and the README notes its macOS support stops around Mojave, citing the project's own release page. For the same subtitle-translation job specifically, the README points to 妙幕 (SmartSub), which generates subtitles from video or audio in batches and supports Baidu, Volcano, OpenAI, Ollama and DeepSeek as translation backends, and to LLM-Subtrans and Subtitle Translator. The meaningful distinction is the backend set: 译幕 routes through conventional translation APIs, while those options include LLM backends, which changes output style and cost structure. The README does not compare them further. ## Maintenance, versioning and what the licence situation means for you The last push to the repository was on 2026-08-28, and the most recent release listed is v5.5.26 on 2026-08-24. Three releases landed within a few days of each other in August 2026, which is consistent with a maintainer shipping fixes in bursts. The README's own commented-out note describes the history as 1.0 in December 2019 through 4.2.0 in February 2021, then a gap until 5.3.0 in November 2024, with an apology and a stated hope for at least one release every three months. That note is commented out, so it is not a current commitment, but it is the only maintenance cadence information the repository offers. The upgrade cost is low by construction. There is no server component you operate and no database to migrate. You download a new build. The risk sits on the other side: provider APIs change, and a pass-through client has to track those changes. The README's speech-to-text removal is the precedent for what happens when a feature outpaces a single maintainer's time. On licensing, the repository metadata does not state a licence. The README does not state one either. That is not a detail you can defer if you plan to redistribute the software, bundle it into a product, or modify it: without an explicit licence, the default position grants you no redistribution rights, and you should treat the terms as unresolved until the repository states them. Nothing here is legal advice, and the practical step is to check the repository for a licence file before you build anything on top of it. The pricing documentation the README links to is the place to confirm what the paid tier currently includes, since the free and paid tier descriptions in the README are largely commented out. ## Conclusion Adopt it if you already hold a translation provider account, or are willing to register one, and you want per-character pricing with no markup and no file uploads to a third party beyond the provider itself. Do not adopt it if you want to drag a file in and get a result without ever visiting a provider console, or if you need speech-to-text, which the README currently comments out. Before committing, verify two things: which providers the current build lists in its dropdown, and whether the target language you need is supported by the provider you picked, since the README states the software only supports what the provider supports. ## FAQ ### How do I translate a subtitle file with 译幕? Configure a translation provider API key first, since the README states the software needs configuration before it can be used. Then drag in a .srt, .ass or .vtt file, pick the provider and the source and target languages, and press start. The configuration walkthrough is linked from the README rather than included in it. ### Is 译幕 a free SRT translator? The software itself does not charge per character, and the README's stated benefit is that you use the provider's free quota, typically one million characters per month. The provider still bills you according to its own terms once that quota is exhausted. The README's free and paid tier descriptions are largely commented out, so check the linked pricing documentation for current terms. ### Can 译幕 convert SRT subtitles to plain text? The README describes 译幕 as a subtitle file translator for .srt, .ass and .vtt, with output as translation only or original plus translation. It does not document a plain-text export. Converting subtitles to text is not a use case the README covers. ### How can I automatically translate subtitles with 译幕? The README describes the flow as dragging a subtitle file in and clicking translate, after a provider is configured. The README does not document a watch folder, batch queue or command-line mode, so automation beyond the manual drag-and-translate step is not described. ### What is the basic workflow for translating a subtitle file in 译幕? Set up a provider first, then drag in the subtitle file, choose the provider and the source and target languages, and click the start button on the right. The README describes the supported input extensions as .ass, .vtt and .srt. ## Sources - [1c7/Translate-Subtitle-File on GitHub](https://github.com/1c7/Translate-Subtitle-File) - [Issues](https://github.com/1c7/Translate-Subtitle-File/issues) - [Project website](https://zimoo.app) - [README](https://github.com/1c7/Translate-Subtitle-File/blob/master/README.md) - [Releases](https://github.com/1c7/Translate-Subtitle-File/releases) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/1c7-translate-subtitle-file