Subtitle Edit

What is Subtitle Edit?

Best suited for

• Independent video editors, creators and small merchants who already have authorized subtitle text and want a reviewable translated SRT track. The local CLI route fits people comfortable running an official utility and supplying an existing model backend. It is particularly useful when shop names, quantities, prices, placeholders and prohibitions must remain visible for bilingual review before a video is published.
• A pilot focused on native srt subtitle translation for creators and small shops with exact timeline and factual boundaries, using an authorized subtitle file with cue numbers, start/end timecodes and text; explicit source and target languages; a chosen translation engine/backend; any protected names, identifiers and factual constraints in a prompt; and a separate output path. The two prepared fixtures contain four UTF-8 SubRip cues each, use English to zh-CN, and provide one exact shared native prompt. They contain fictional shop text rather than customer records or a real order.

Not suited for

• Use without the inputs, access and review described in the pilot dependencies.
• The full Subtitle Edit GUI and official SeConv are distinct interfaces. The planned pilot covers only the complete native SeConv v5.2.0 Windows x64 SRT translation path; it cannot establish desktop interaction, speech recognition, dubbing, burn-in, upload or publishing behavior.
• Translation depends on the selected backend and model. A small coding-oriented Qwen model is not a product-wide translation benchmark; its results would apply only to the recorded model, prompts and runtime.

Capabilities, with sources

• 01Official documentation describes Subtitle Edit 5 as a free, open-source editor for video subtitles and lists auto-translation as a separately documented feature.Official vendor statement · checked 2026-10-03Source ↗
• 02The official documentation separately lists AI Assistant, AI Review, speech-to-text and text-to-speech; listing these features does not establish that the prepared SeConv subtitle cases test them.Official vendor statement · checked 2026-10-03Source ↗
• 03Official SeConv is a headless command-line converter sharing Subtitle Edit core libraries without a desktop GUI dependency.Official vendor statement · checked 2026-10-03Source ↗
• 04The fixed command-line reference documents source/target languages, translation engine, an existing backend URL, a custom prompt and explicit output-folder/output-filename options.Official vendor statement · checked 2026-10-03Source ↗
• 05The fixed native translate runner uses an explicitly supplied llama.cpp URL directly, providing a route to an already-running backend without its local model/server startup path.Official vendor statement · checked 2026-10-03Source ↗
• 06The fixed LlamaCppTranslate implementation sends a chat-completions request, omits the model field, conditionally omits default sampling fields and performs native response-text cleanup.Official vendor statement · checked 2026-10-03Source ↗
• 07The fixed translation loop contains product-internal retries and line fallback; one native conversion process can produce more than one backend request.Official vendor statement · checked 2026-10-03Source ↗
• 08The fixed v5.2.0 release exposes an official SeConv-Windows-x64.zip distribution; the build workflow publishes self-contained binaries and copies the license.Official vendor statement · checked 2026-10-03Source ↗
• 09The v5.2.0 MIT license names Nikolaj Olsson as copyright holder; the official niksedk GitHub User profile shows the same personal name.Official vendor statement · checked 2026-10-03Source ↗
• 10The retained official README describes offline core functions and says optional online translation and other providers receive request data under the selected provider privacy policy; this is the project statement, not a full independent privacy audit.Official vendor statement · checked 2026-10-03Source ↗

Inputs and outputs

Inputs

An authorized subtitle file with cue numbers, start/end timecodes and text; explicit source and target languages; a chosen translation engine/backend; any protected names, identifiers and factual constraints in a prompt; and a separate output path. The two prepared fixtures contain four UTF-8 SubRip cues each, use English to zh-CN, and provide one exact shared native prompt. They contain fictional shop text rather than customer records or a real order.

Outputs

A native converted subtitle file in the selected format, with translation text and a conversion result examined separately from model responses. Both original cases saved their first nonempty native UTF-8-with-BOM SubRip file, with all four original cue numbers and exact corresponding start/end millisecond timecodes intact. The first native subtitles still contained the translation errors reported in Tests; file creation and timeline preservation did not establish semantic accuracy. Native text cleanup, grouping and splitting are product behavior, so raw provider replies are retained separately and are not substituted for the native files.

A practical Subtitle Edit workflow

1. Prepare the native srt subtitle translation for creators and small shops with exact timeline and factual boundaries fixture: 1 00:00:00,000 --> 00:00:03,000 Welcome to Juniper Paper. 2 00:00:03,500 --> 00:00:07,000 We need 3 notebooks and 2 pens. 3 00:00:07,500 --> 00:00:10,500 The total budget is 12 dollars. 4 00:00:11,000 --> 00:00:15,000 Do not place an order. Pickup is Friday at 4 in the afternoon.
2. Check Subtitle Edit access through Desktop subtitle editor, Official headless SeConv CLI, Bring your own translation backend and confirm the selected feature’s actual permissions.
3. Translate from {0} to {1}. Keep all line breaks and punctuation as input. Keep brand and shop names, bracketed identifiers such as [ORDER_ID], and Arabic numerals unchanged. Preserve negation and statements that a value is unknown. Output only the translation, without comments:
4. Inspect one actual native Chinese SubRip subtitle file, retaining all four original cue numbers, exact start/end millisecond times, shop name Juniper Paper, quantities 3 notebooks and 2 pens, total budget 12 dollars, prohibition on placing an order, and Friday at 4 in the afternoon pickup. No new claims or commentary. Human/bilingual semantic assessment follows the frozen complete cue-by-cue rules; no selective token-region deletion. Compare it against the source input and retain the output/action log.
5. Run the boundary case: 1 00:00:00,000 --> 00:00:04,000 [ORDER_ID] is a placeholder, not a real order. 2 00:00:04,500 --> 00:00:08,000 みどり文具: prepare 2 drafts for review only. 3 00:00:08,500 --> 00:00:12,000 The pickup date and profit are unknown. 4 00:00:12,500 --> 00:00:17,000 Do not send an email or promise a refund. Accept the result only if all pass conditions are met and no failure condition occurs.

This is an evaluation workflow built around the documented product scope. Check feature and plan eligibility before expecting the vendor product to complete every step.

Setup and integrations

Subtitle Edit 5 is documented as a cross-platform desktop application. SeConv is the official headless utility that reuses libse/libuilogic without a GUI dependency. The recorded component was the complete SeConv-Windows-x64.zip v5.2.0 release asset at source commit d8e3b8b41e856a896c541ce7e59b490a99c21196. Its official build publishes a self-contained Windows distribution and copies the MIT license. The complete ZIP, extracted native executable, distribution inventory, original fixtures/prompt, existing model and runtime were bound in the pre-inference manifest. Each official native process exited 0 and saved its first SRT; both cases nevertheless failed their original semantic conditions. These component/runtime checks do not establish desktop GUI or full product-quality acceptance.. Documented access methods: Desktop subtitle editor, Official headless SeConv CLI, Bring your own translation backend. Confirm each method’s plan eligibility and actual action scopes before connecting an account.

Access and setup steps

1. Start with an authorized subtitle track and keep the original as a separate file. Confirm source/target language and protected shop names, identifiers, numbers, currencies, negations and unknown values before translating.
2. Choose the desktop editor or official headless SeConv according to the task. For this reproducible local pilot use the complete official v5.2.0 Windows x64 SeConv ZIP; verify the release-asset digest, actual extracted distribution inventory and native executable hash without substituting another converter.
3. Run native help, format listing and dump-settings before inference, and retain stdout/stderr/exit codes and effective settings. Record actual paths, arguments, backend/model identity and original fixture/prompt bytes in a pre-inference manifest.
4. Point the native llamacpp translation engine at the already-running authorized backend with an explicit translate-url. This takes the existing-server route; verify the selected endpoint and do not trigger a new model download, paid provider or unrelated backend configuration change.
5. Supply explicit English and zh-CN languages, SubRip output, the frozen native prompt file and a fresh output folder/name. The native prompt substitutes its language placeholders; the original case input remains the exact SRT. Keep each case separate with one native conversion start and its fixed request/timeout budget.
6. Retain actual sampling omission, computed max_tokens, every product-internal grouping/retry/fallback request and complete provider response. Do not infer an omitted sampling value as zero or replace native settings through the recorder.
7. Review the first saved native SRT for exact cue numbers/timecodes, complete Chinese meaning, literal protected names, quantities, dollars, negations and unknowns. Keep process errors or missing output visible rather than fixing the subtitle to improve a score.
8. Save the original input, native output, raw backend exchanges, process outcome and independent condition review. Confirm reading speed and final video synchronization separately before using the subtitles commercially or publishing the video.

Test access: local install. Both original cases completed once on 2026-10-03 through the complete official SeConv-Windows-x64.zip v5.2.0 native CLI subcomponent, not a desktop GUI or substitute parser. Each used the original four-cue UTF-8 SRT, English to zh-CN, the unchanged shared native prompt, an explicit already-running llama.cpp URL and a fresh native output path. Actual distribution/native help/settings/model identity and the complete case contract were frozen before inference. Both native processes exited 0 and saved first UTF-8-with-BOM SRTs with exact original cue/timecode mapping; both original cases failed, with primary 4/6 and boundary 3/6 conditions passed, five failures total and zero unverified. Two native starts and two provider forwards were recorded, with no observed native fallback, quality rerun or manual repair. Native payloads omitted model, stream and sampling fields; max_tokens was 512 for primary and 544 for boundary. Retained backend defaults were temperature 0.800000011920929, top_p 0.949999988079071, top_k 40, repeat_penalty 1 and seed sentinel 4294967295, distinct from explicit request settings; the actual sampled seed is unknown. Native response cleanup/grouping/splitting and first files are preserved alongside raw provider replies. GUI, ASR, dubbing, OCR, burn-in, synchronization, upload and publication remain untested. Open the official access or installation page ↗

Pilot dependencies

• Complete official SeConv v5.2.0 Windows x64 distribution at source commit d8e3b8b41e856a896c541ce7e59b490a99c21196, verified original fixture/prompt bytes, effective native help/settings and pre-inference runtime manifest. Use the already-authorized cached Qwen2.5-Coder-1.5B-Instruct Q4_K_M backend through a fresh unchanged-payload recorder; explicit native llamacpp translate-url, en to zh-CN, one native process per case, 180 seconds and maximum twelve forwards per case. Retain every product-internal retry/fallback, first saved native SRT and process/request/response hashes. No new model, paid provider, external business action, quality rerun, GUI, ASR, burn-in or publication belongs to this translation plan.
• Both original cases completed once on 2026-10-03 through the complete official SeConv-Windows-x64.zip v5.2.0 native CLI subcomponent, not a desktop GUI or substitute parser. Each used the original four-cue UTF-8 SRT, English to zh-CN, the unchanged shared native prompt, an explicit already-running llama.cpp URL and a fresh native output path. Actual distribution/native help/settings/model identity and the complete case contract were frozen before inference. Both native processes exited 0 and saved first UTF-8-with-BOM SRTs with exact original cue/timecode mapping; both original cases failed, with primary 4/6 and boundary 3/6 conditions passed, five failures total and zero unverified. Two native starts and two provider forwards were recorded, with no observed native fallback, quality rerun or manual repair. Native payloads omitted model, stream and sampling fields; max_tokens was 512 for primary and 544 for boundary. Retained backend defaults were temperature 0.800000011920929, top_p 0.949999988079071, top_k 40, repeat_penalty 1 and seed sentinel 4294967295, distinct from explicit request settings; the actual sampled seed is unknown. Native response cleanup/grouping/splitting and first files are preserved alongside raw provider replies. GUI, ASR, dubbing, OCR, burn-in, synchronization, upload and publication remain untested.
• Confirm free mit application · model, service and hardware costs separate against the current vendor terms; usage and connected-service costs can affect the pilot.
• Create a test workspace or use public/authorized material. Keep an input baseline, output artifact and action log for comparison.

Named native platform connections have not been verified in this profile.

Content output describes an export suited to a channel; marketplace data describes research coverage. Exact data scopes and permissions need a setup review.

API: Not verifiedNot verified in the reviewed official material.

Self-hosting: Yes (documented)Documented local/self-hosted option; model inference, license and infrastructure conditions require separate review.Source 1

Open source: Yes (documented)The official source names a conventional open-source license; verify the license of the exact distribution and related services.Source 1

Pricing and additional costs

Free MIT application · model, service and hardware costs separate

The official documentation describes Subtitle Edit as free and open-source, and the fixed v5.2.0 application license is MIT. The app license does not include a language model, cloud tokens, speech-recognition models, GPU, memory or electricity. Optional online providers have their own access and fees; local translation needs a separately licensed model and compatible backend. The prepared route reuses an already cached Qwen model and an existing llama.cpp server, with no new model or paid provider proposed. That selected setup does not establish a free total operating cost, included cloud entitlement or unlimited translation.

Free application licensing is separate from inference, optional engine licensing, hardware and any hosted provider tariff. Check the exact chosen backend and distribution before committing to a production subtitle workflow; no paid-provider price or included allowance was verified for these cases.

Budget for the base plan, usage limits, connected services, licensing, implementation and human review where applicable.

Test plan and results

The cases below define what to supply, what to inspect and what would pass. A planned case is not a completed product test.

See the testing method and all product plans →

Dependencies before a product pilot

• Complete official SeConv v5.2.0 Windows x64 distribution at source commit d8e3b8b41e856a896c541ce7e59b490a99c21196, verified original fixture/prompt bytes, effective native help/settings and pre-inference runtime manifest. Use the already-authorized cached Qwen2.5-Coder-1.5B-Instruct Q4_K_M backend through a fresh unchanged-payload recorder; explicit native llamacpp translate-url, en to zh-CN, one native process per case, 180 seconds and maximum twelve forwards per case. Retain every product-internal retry/fallback, first saved native SRT and process/request/response hashes. No new model, paid provider, external business action, quality rerun, GUI, ASR, burn-in or publication belongs to this translation plan.
• Both original cases completed once on 2026-10-03 through the complete official SeConv-Windows-x64.zip v5.2.0 native CLI subcomponent, not a desktop GUI or substitute parser. Each used the original four-cue UTF-8 SRT, English to zh-CN, the unchanged shared native prompt, an explicit already-running llama.cpp URL and a fresh native output path. Actual distribution/native help/settings/model identity and the complete case contract were frozen before inference. Both native processes exited 0 and saved first UTF-8-with-BOM SRTs with exact original cue/timecode mapping; both original cases failed, with primary 4/6 and boundary 3/6 conditions passed, five failures total and zero unverified. Two native starts and two provider forwards were recorded, with no observed native fallback, quality rerun or manual repair. Native payloads omitted model, stream and sampling fields; max_tokens was 512 for primary and 544 for boundary. Retained backend defaults were temperature 0.800000011920929, top_p 0.949999988079071, top_k 40, repeat_penalty 1 and seed sentinel 4294967295, distinct from explicit request settings; the actual sampled seed is unknown. Native response cleanup/grouping/splitting and first files are preserved alongside raw provider replies. GUI, ASR, dubbing, OCR, burn-in, synchronization, upload and publication remain untested.
• Confirm free mit application · model, service and hardware costs separate against the current vendor terms; usage and connected-service costs can affect the pilot.
• Create a test workspace or use public/authorized material. Keep an input baseline, output artifact and action log for comparison.

Official-page checks

Evidence

Official documentation

Claims cited on this page, with source access status below. URL accessibility is separate from a substantive claim review.

Public feature checks

No public feature output or demonstration has been independently assessed for this profile.

uAgentKit product execution

Local model product test · 2 cases executed. 2 of 2 defined cases have actual execution records; their outcomes, access method and disclosed execution metadata appear in the test section. Official complete SeConv v5.2.0 Windows x64; English to Chinese (Simplified) subtitle translation preserving first native UTF-8 SubRip files for two frozen fictional inputs. Both native conversions completed, but both original cases failed with the recorded local model: 7 of 12 conditions passed and 5 failed. Timecodes were retained; first subtitles contain mistranslated stationery/pickup content or changed protected identifiers/numerals. Conditions and first outputs were not repaired.

uAgentKit website acceptance

Visible profile structure and content checks are reported in the test section; these evaluate this directory page.

Professional review

Not conducted by a clinician, lawyer, agronomist, investment professional or security auditor.

Commercial use: The fixed MIT application license allows commercial use subject to retaining its copyright and permission notice; it offers no warranty. It does not automatically grant rights to caption text, source videos, names, third-party components, model weights or provider outputs. Check the selected backend terms and review factual and translation content before commercial use.

Limitations and checks

• The full Subtitle Edit GUI and official SeConv are distinct interfaces. The planned pilot covers only the complete native SeConv v5.2.0 Windows x64 SRT translation path; it cannot establish desktop interaction, speech recognition, dubbing, burn-in, upload or publishing behavior.
• Translation depends on the selected backend and model. A small coding-oriented Qwen model is not a product-wide translation benchmark; its results would apply only to the recorded model, prompts and runtime.
• Cue counts and timestamps can remain syntactically valid while text loses a quantity, negation, name or unknown fact. Review the complete saved SRT against every original cue, then check it against the intended video separately.
• Native text cleanup, grouping, merge/split, retries and single-line fallback are part of the product. Preserve all actual requests and the first native file; do not replace a failed result with a hand-corrected subtitle or a preferred raw model reply.
• Local core editing does not prove every optional feature stays offline. A configured online translation, speech, OCR or other provider can receive supplied data under its own terms; the prepared localhost route is narrower than a full privacy audit.
• The prepared cases translate instructions that forbid email/refund/order actions; they do not test autonomous tool permission or model-initiated refusal. No real business account or external action tool belongs to their input.
• A published release record and verified source claims do not establish downloaded ZIP integrity, native startup, file-save success or translation quality. These must be recorded in runtime and first-output evidence.
• The official main README was retained on 2026-10-03; component and case mechanics use the fixed v5.2.0 source. Do not treat future main changes or another release as the same tested configuration.
• Current team size, legal operating entity and ownership remain unknown. A personal copyright notice or sponsor link does not prove a present-day one-person company or a lack of acquisition.
• Both original first native subtitle cases failed with the recorded cached Qwen model: primary passed four of six conditions and boundary three of six. Pens became pencils and pickup became arrival; [ORDER_ID] and みどり文具 were replaced, Arabic 2 became 两, and placeholder remained English. Exact cues/timecodes, unknown pickup/profit and prohibitions were retained. Seven of twelve conditions passed, including workflow/structure/evidence conditions; this is not translation accuracy or a product-wide rating. No corrected subtitle or quality rerun replaces the first result.

Field-level unknowns identify gaps in this review. They do not imply the vendor lacks the capability.

Alternatives and comparisons

No editorial comparison or alternative guide meets the publication standard for this product yet. Build an instant fact comparison.

Questions about Subtitle Edit

Sources and change history

1. Official Subtitle Edit 5 documentation: subtitle editor and separately documented AI features

   Subtitle Edit project / Nikolaj Olsson · subtitleedit.github.io · Read · 2026-10-03

2. Official retained main README: platforms, optional-provider data paths and sponsor link

   Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

3. Fixed v5.2.0 SeConv README: official headless converter using Subtitle Edit core libraries

   Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

4. Fixed v5.2.0 native command-line reference: subtitle translation and explicit output controls

   Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

5. Fixed v5.2.0 MIT license and Nikolaj Olsson copyright attribution

   Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

6. Official GitHub niksedk profile: Nikolaj Olsson individual account attribution

   Subtitle Edit project / Nikolaj Olsson · api.github.com · Read · 2026-10-03

7. Fixed v5.2.0 application About source: Subtitle Edit links and niksedk sponsor attribution

   Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

8. Official observed v5.2.0 release record and Windows x64 SeConv distribution asset

   Subtitle Edit project / Nikolaj Olsson · api.github.com · Read · 2026-10-03

9. Official v5.2.0 tag commit and source tree identity

   Subtitle Edit project / Nikolaj Olsson · api.github.com · Read · 2026-10-03

10. Fixed v5.2.0 official SeConv self-contained distribution build and copied license

    Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

11. Fixed v5.2.0 native translate runner and already-running llama.cpp URL route

    Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

12. Fixed v5.2.0 native llama.cpp payload, omission defaults and response cleanup

    Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

13. Fixed v5.2.0 official translation loop, built-in retries and line fallback

    Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

14. Fixed v5.2.0 native text grouping and subtitle merge/split helper

    Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

15. Fixed v5.2.0 library translation defaults and sampling sentinel settings

    Subtitle Edit project / Nikolaj Olsson · raw.githubusercontent.com · Read · 2026-10-03

Suggest a sourced correction →