filmcraft/docs/target-app-parity.md

Parity with Adobe Premiere Pro

Last reviewed: 2026-10-10 · Last updated: 2026-10-10 · Change: major (full re-measure against Premiere Pro 2026 26.5.2; replaces the scorecard and estimate that lived in ROADMAP.md) · Target: Adobe Premiere Pro 2026 (26.5.2, macOS)

The authoritative assessment of how close FilmCraft is to Premiere Pro, and how much work is left. ROADMAP.md summarizes it; gaps.md lists every shortfall one at a time. The deep checklists are file-format-parity.md, codec-parity.md, hardware-parity.md, ui-parity.md and localization-parity.md.

Headline

	Value	Kind
Feature breadth	~86%	Partly measured: menus 92% (label diff, below), effects / transitions / audio effects 100% (unit tests); the other areas are estimated
Ready for real work	~56% (range 51–61%)	Estimated depth per area, combined with written weights (formula)
Mainstream practitioner	~55%	Weekly editing areas only, with written discounts (method)
Essentials user	~60%	Core features at default settings, with written discounts (method)
Stage	alpha	~19 points and ~400–700 h from beta, and beta also needs .prproj import (why)
Remaining to beta	~400–700 Opus 5.5 agent-hours	Estimated, calibrated below
Remaining to full parity	~1,550–2,750 Opus 5.5 agent-hours (to ~95%: ~1,450–2,600)	Estimated; ~70% parallelizes
Codebase (measured 2026-10-10)	45 crates + 3 apps, ~343,000 lines of Rust, 2,520 #[test] functions, 1,098 commits, 268 merged PRs, 69 authors	git, gh, grep

The previous estimate (2026-10-05) was breadth ~87%, ready ~50–60%, and "~210–325 agent-hours to 100%". The breadth figure is one point lower now because this pass counts the file formats Premiere lists and we lack (AVI, WMV, camera RAW, EXR / DPX / PSD stills, .prproj, DVD / Blu-ray / DCP export) rather than only the formats we have. The hours went up because the old figure covered the checklist only and left out plugins, AI, localization, .prproj and the depth work that 188 open issues now document. That is new evidence, not drift: see Calibration.

What we measured against, and how

• Target: Adobe Premiere Pro 2026, version 26.5.2 (Spotlight kMDItemVersion of the installed /Applications/Adobe Premiere Pro 2026), plus Adobe Media Encoder 2026 for export.
• Menus (measured). The maintainers' native menu-bar dump of Premiere 26.5.2 (plan/premiere/menus.json, taken 2026-09-30 through System Events; local only, never committed) has 433 leaf items. Removing Adobe-cloud, account, help, macOS system and recent-file entries leaves 375. A label match against the string literals of crates/engine, crates/ui-egui, crates/edit and crates/project finds 338 of 375 (90.1%); of the 37 misses, 9 are dynamic or system items (About, AutoFill, Dictation, Emoji, window names), so in-scope coverage is 338 / 366 ≈ 92%. Real misses: Spelling (2), Render and Replace / Restore Unrendered, Enhance Speech toggle, Auto-Tag Audio Types, Auto Reframe Sequence, Generate / Reassociate Source Clips, Project Shortcut, Refresh All Projects, Export Selection as Premiere Project, 5 workspaces (Essentials, Metalogging, Prelude, Starter, Text-Based Editing), Audio Clip / Track Effect Editor, Sequence Index, AI Models and Collaboration settings, UXP / Exchange plug-in entries. A label match is presence only, and can over-count (a label that exists in a different place); it is a lower bound on misses, not proof of behaviour. The script was ad hoc; cargo xtask parity now measures the shipped menu bar against the snapshot, reporting presence and placement separately (not yet re-run for this figure); the other areas are gap G7.
• Effects (measured by tests). premiere_26_video_effects_catalogue, premiere_audio_set_is_complete_and_foldered and premiere_26_transition_tree check the catalogue against Premiere's Effects panel: video effects 93/93 (+Legacy, Obsolete), audio effects 53/53, video transitions 84/84 (+21 Legacy), audio transitions 3/3. cargo xtask parity --area effects|transitions measures the shipped catalogue against a reference list too (presence and folder placement, G7).
• The installed app (file names and listings only, per AGENTS.md §2; nothing inside the bundle was opened or read). Contents/Resources/*.lproj: de, en, es, fr, it, ja, ko, pt, ru, zh_CN (10 UI languages). Contents/MediaIO/systempresets/ folder names are four-character exporter / codec codes: JPEG, QuickTime, MP3, PNG, Windows Media, AIFF, AVI, BMP, DPX, TIFF, Targa, WAV, HEVC, MPEG-2 / DVD / Blu-ray, MXF (JPEG 2000, DNx, P2, XDCAM HD, AS-10, AS-11), MP4 / AAC / H.264 / H.264 Blu-ray, FLV, uncompressed AVI, DCP, GIF, raw PCM, OpenEXR. Contents/Frameworks names show what it decodes natively: ARRIRAW (ARRI Image SDK), RED R3D, Sony RAW, Canon RAW (CrmSdk), Codex HDE, DNx SDK, JPEG XS, JPEG 2000 (Kakadu), DV100, MPEG-4 Part 2, OMF (Pro4OMF), AAF, LTC timecode, C2PA content credentials, and an ONNX runtime for its AI features. PlugIns lists OpenEXR, Radiance and RED / MXF XMP handlers.
• Our side: this repository on origin/main at 98ba9ac (2026-10-10): code, crate READMEs, tests, docs/performance.md benchmarks, and the 188 open / 118 closed GitHub issues (user reports from releases 0.2–0.5).
• Public documentation: Adobe's "What's new" and release notes for 26.0–26.3 (Object Mask, redesigned mask tools, R3D NE, OpenTimelineIO, GPU thumbnails, 4:2:2 10-bit hardware decode on NVIDIA Blackwell), and Adobe's published lists of supported file formats and GPU / hardware acceleration.
• Not done in this pass: running Premiere side by side on shared projects. No rendering or behaviour of ours is compared with Premiere's output automatically yet (G1).

Weights

Feature areas carry 70% of "ready for real work", cross-cutting qualities 30%. Weights reflect what a working editor uses daily: the timeline first, then media, effects and panels.

Group	Weight
Editing, timeline, trimming, multicam	20% of features
Effects and transitions	12%
Media import and codecs	12%
Audio	10%
Panels and UI fidelity	10%
Colour	8%
Graphics, titles, captions	8%
Export and delivery	8%
Project files and interchange	8%
Preferences and project management	4%
Cross-cutting: hardware and performance	10% of overall
Stability	8%
Platforms	5%
Plugins and ecosystem	5%
AI features	2%

Feature breadth weights the same ten feature areas only (cross-cutting qualities have no "exists / doesn't" count).

By feature area

Breadth = does it exist. Ready = depth, correctness, fidelity, robustness. Hours = Opus 5.5 agent-hours to close the area to Premiere, sequential.

Area	Weight	Breadth	Ready	Hours	Evidence	Detail
Editing, timeline, trimming, multicam	20%	~90%	~70%	80–140	Edit algebra, trim modes, Trim Monitor, multicam with audio sync, Premiere default keyboard (143 commands), Source playback and range dragging (10-08). Open: ripple-delete of a gap by column (#668), Backspace on gaps (#648), clip stretching (#653), deselect on empty space (#683), new tracks by dragging media (#483), Escape cancelling drags (#580 fixed), time remapping (#510), effects on a whole track (#623)	ui-parity.md
Effects and transitions	12%	100% (measured)	~60%	80–140	Every effect exists; Warp Stabilizer is 2-D, Auto Reframe approximate, Morph Cut morphs with optical flow but without face tracking (#805); Optical Flow time interpolation is motion compensated on the CPU (#780). 34 effect ids run on the GPU against ~82 of Premiere's ~92 marked Accelerated. No Object Mask (26.0, AI).	hardware-parity.md
Media import and codecs	12%	~75%	~60%	150–250	Own bit-exact decoders for H.264, HEVC, VP9, AV1, ProRes, DNx, APV, MPEG-2, AAC, AC-3, Opus. But H.264 only 8-bit 4:2:0 and HEVC no 4:2:2 in software (most 10-bit cameras, #626); no camera RAW (R3D, ARRIRAW, BRAW, Sony RAW, Canon RAW, ProRes RAW, #345, #383); no E-AC-3 (#647), AVI (#598), WMV, DV, XAVC / P2 metadata; Sony start timecode (#460); WebM black (#432)	codec-parity.md, file-format-parity.md
Audio	10%	~88%	~65%	50–90	Mixer graph, automation, all 53 effects, Essential Sound, 5.1, voice-over, Remix, multi-stream container audio (#361). Device and playback reports: crackle / cut-outs (#336, #410, #602, #605, #617); Enhance Speech is DSP, not a model; ALSA fallback fixed (#254)
Panels and UI fidelity	10%	~88%	~60%	120–200	All main panels. Missing: on-screen transform handles in the Program monitor (#639), Effect Controls keyframe zoom / divider / row lines (#640–#645), docking rearrangement (#493), multi-monitor windows (#466), filmstrip thumbnails (#614), marquee in Project panel (#578)	ui-parity.md
Colour	8%	~90%	~70%	40–70	Lumetri complete incl. scopes, LUTs, colour management, HDR export. Lumetri on the GPU covers basic, creative and vignette only; HDR-aware Lumetri maths, macOS EDR display, D-Log M; a Lumetri report (#628)
Graphics, titles, captions	8%	~85%	~65%	50–90	Text engine with shaping and bidi, shapes, gradients (#514), our .fcgt templates (.mogrt refused by design), rolls / crawls, all caption formats. Arabic titles reported broken (#395); no 608/708 embedding in the video stream; speech-to-text off in release builds
Export and delivery	8%	~80%	~65%	60–110	H.264 (own encoder), HEVC through hardware only, ProRes, DNxHR, APV, MXF OP1a / OP-Atom, image sequences, GIF, WAV / AIFF, presets, queue, GPU export rendering (#421). Missing: AV1, MPEG-2 / DVD / Blu-ray, DCP, JPEG 2000 / XS MXF, AS-10 / AS-11, XDCAM HD, EXR / DPX / Targa sequences, MP3, AAC-only audio, ProRes 4444 in the UI (#342), interlaced, smart render; bitrate target missed (#371)	file-format-parity.md
Project files and interchange	8%	~70%	~45%	90–160	.fcproj with migrations, auto-save, recovery; FCP7 XML, FCPXML, AAF, OMF, EDL, OTIO, ALE. Premiere's own .prproj cannot be opened (the main beta blocker); AAF / OMF never validated in Media Composer / Pro Tools; XML import of a Premiere sequence took 97 s (#461)	file-format-parity.md
Preferences and project management	4%	~80%	~65%	20–40	16 Settings categories, most wired; Project Manager, proxies, Link Media; Media Manager limits (#411); images as proxies (#451)
Features, weighted	100%	~86%	~64%	740–1,290

By dimension

Dimension	Ready	Hours	Doc
Features (editing, effects, colour, audio, graphics, export settings, preferences)	~66%	350–620	this file
UI / UX fidelity	~60%	120–200	ui-parity.md
File formats (containers, project, interchange, captions, stills)	~60%	140–240	file-format-parity.md
Codecs (decode and encode, fidelity, camera media)	~60%	150–250	codec-parity.md
Hardware acceleration (decode, encode, GPU effects, I/O devices)	~45%	120–200	hardware-parity.md
Localization (12 key languages)	~35% of Premiere's set; 4 of 12 key languages near-complete	150–230	localization-parity.md
Performance (beyond hardware)	~55%	60–120	performance.md
Stability	~45%	60–120	gaps.md
Platforms	~60%	50–90	gaps.md
Plugins and ecosystem	~5%	200–400	gaps.md
AI features	~25%	150–300	gaps.md
Total	~56%	~1,550–2,750

Dimension hours are disjoint (the feature row excludes UI, formats and codecs), so they add up. Evidence for the cross-cutting rows:

• Hardware ~45%. Hardware decode exists on all three desktop OSs (VideoToolbox; Media Foundation / D3D11 incl. VP9 and AV1; VA-API and NVDEC for H.264 / HEVC on Linux). Hardware encode: VideoToolbox H.264 / HEVC 8-bit; NVENC H.264 (Windows, Linux) and HEVC incl. Main 10 HDR (Windows). Missing against Premiere: Intel Quick Sync and AMD encoders, 10-bit HEVC on macOS, ProRes hardware, zero-copy upload, ~48 of Premiere's accelerated effects, external video I/O, control surfaces. Detail: hardware-parity.md.
• Performance ~55%. 4K H.264 / HEVC play with no dropped frames with hardware decoding (M4 Pro); GPU export is 2.4–3.2× faster on effects-heavy timelines; AV1 software decode is still slow and 8K is not real time. Users on 0.5.0 report constant lag and ~100% CPU (#523, #525) and a CPU fallback adapter rendering everything in software (#632).
• Stability ~45%. Never-crash rules enforced by lints; hostile-input fuzzing. But open launch crashes on macOS 12 (missing VideoToolbox symbol, #468 / #661, #655), Intel UHD graphics (#512), Windows 11 (#380, #582), a generic "keeps crashing" (#687), and no workflow runs the test suite on pull requests (.github/workflows has release, packaging lint, FreeBSD and Windows ARM64 builds only); cargo xtask ci runs locally only.
• Platforms ~60%. Releases for macOS, Windows (x64, ARM64), Linux (AppImage, Flatpak, deb, tar) and FreeBSD; Linux is something Premiere doesn't run on at all. Runtime testing is still mostly macOS; Windows and Linux reports dominate the issue list. The web build runs but its shell is incomplete. No mobile.
• Plugins ~5%. No VST3 / Audio Units, no OpenFX, no Premiere / After Effects plug-in API, no UXP panels (#325). Needs an owner decision (FFI). MCP, CLI and the control channel are ahead of Premiere's scripting.
• AI ~25%. Speech to text exists (local Whisper) but is off in release builds; text-to-speech narration with Kokoro-82M landed 10-10 (#278, beyond Premiere); scene edit detection; Enhance Speech and Auto Reframe are DSP / heuristics. Missing: Object Mask (#406, #694), Generative Extend, media-intelligence search, caption translation, auto colour.

Three readiness numbers

Audience	Ready %	Opus 5.5 agent wall-clock hours to ~95%	Work that dominates
Full target (ready for real work)	~56%	~1,450–2,600 (~70% parallelizes)	Plugin hosting, AI models, localization incl. RTL / Devanagari, camera RAW and delivery codecs, .prproj, hardware acceleration, UI fidelity
Mainstream practitioner	~55%	~500–850 (~65% parallelizes)	.prproj import (60–120), UI fidelity (120–200), stability and PR CI (40–80), 10-bit / 4:2:2 camera decode (30–50), playback / hardware (60–100), import and export gaps (60–100)
Essentials user	~60%	~180–350 (~60% parallelizes)	Launch crashes on Windows / Intel / older macOS (40–80), discoverability and timeline friction (60–110), playback lag (30–60), common import gaps (20–40), audio devices (15–30)

Hours per audience are calibrated like the rest (one hardware backend ~3–5 h, one encoder ~4–8 h, a whole-interface language ~4–8 h, ~4–5 h per breadth point, depth work ~1.5–2× that). The audiences nest, so essentials ≤ mainstream ≤ full: the essentials work is a subset of the mainstream list, which is a subset of the full list. Needs a human: the plugin decision and AI models (full only), native-speaker review (full only), Avid / Pro Tools validation (mainstream and full), Windows / Intel test machines (all three).

All three use the depth ("Ready") estimates of the feature areas; they differ in which areas count and how they are discounted. The stage follows the first one and the core-workflow gate.

Ready for real work (full target)

Written formula: 0.70 × (feature areas, weighted as in Weights) + the cross-cutting rows, each weight × its percent.

• Feature areas: 0.20×70 + 0.12×60 + 0.12×60 + 0.10×65 + 0.10×60 + 0.08×70 + 0.08×65 + 0.08×65
  • 0.08×45 + 0.04×65 = 63.1%, × 0.70 = 44.2 points.
• Cross-cutting: hardware and performance 10% × 45, stability 8% × 45, platforms 5% × 60, plugins 5% × 5, AI 2% × 25 = 4.5 + 3.6 + 3.0 + 0.25 + 0.5 = 11.9 points.
• Total 56.0% ≈ ~56% (range 51–61%). Earlier in this pass the headline said ~55% (range 50–60%): the same inputs, rounded down by judgement instead of computed. Nothing in the evidence changed; the number moved only because the arithmetic is now written out. The stage is unaffected (still alpha; beta needs ~75% and .prproj).

Mainstream practitioner

A typical professional editor's week: cut, trim, colour, mix, title, caption and deliver, on their own machine, exchanging projects with Premiere users. Excluded: plugins, generative and cloud AI, Team Projects / Productions, control surfaces and external I/O, other interface languages. Depth here is judged on what that editor actually uses inside each area, so it is higher than the full-area depth (e.g. the ~20 common effects rather than all 93 and their approximations).

Area	Weight	Depth	Why
Editing and trimming	28%	80%	Insert / overwrite, ripple / roll / slip / slide, razor, J/K/L, markers, multicam all work; gap deletion (#648, #668) and deselect (#683) are friction
Importing everyday footage	14%	65%	Phone and 8-bit camera H.264 / HEVC, ProRes, DNx, MXF open; 10-bit / 4:2:2 H.264 (Sony XAVC S-I, Canon) does not (#626); E-AC-3, AVI missing
Common effects and transitions	9%	75%	Lumetri, Transform, Crop, blurs, Cross Dissolve, keys exist; Warp Stabilizer is 2-D, Object Mask missing
Colour (Lumetri, scopes, LUTs)	9%	80%	Complete, parity-tested CPU / GPU; curves / wheels on CPU
Audio (levels, Essential Sound, loudness, ducking)	11%	75%	Complete; audio-device reports off macOS
Titles and captions	7%	75%	Text, shapes, templates, SRT / SCC / VTT; .mogrt not read by design
Export with presets	9%	75%	H.264, HEVC (hardware), ProRes, DNxHR, presets, queue; bitrate miss (#371)
Save, reopen, auto-save, relink, proxies	4%	80%	Atomic saves, migrations, recovery, Link Media
Playback speed on their machine	9%	65%	4K H.264 / HEVC real time with hardware decode on macOS; lag / CPU reports (#523, #525)
Weighted depth	100%	74.8%

Discounts for what still stops real work:

• Interaction fidelity ×0.90: about 40 open UI issues; no transform handles in the Program monitor (#639); Effect Controls keyframe zoom / divider (#640–#645); docking (#493).
• Stability on real machines ×0.92: current macOS on Apple silicon is solid, but launch crashes on macOS 12, Intel UHD and Windows 11 (#468, #512, #380, #582, #687), and no CI runs the tests on pull requests.
• File exchange with Premiere users ×0.88: .prproj does not open (colleagues send projects, not XML); AAF / OMF never validated in Avid / Pro Tools; FCP7 XML / OTIO do work.

74.8 × 0.90 × 0.92 × 0.88 = 54.5% ≈ ~55%. It is not higher than the full number, unlike most Craft apps: the parts the mainstream number leaves out (plugins, AI, localization) weigh only ~12% in the full number, while the three discounts hit exactly what a working editor meets every week. Raising it means fixing G2, G3, G4 and G8.

Essentials user

Someone who imports phone or camera clips, puts them on a timeline, trims, adds a title, music and a dissolve, and exports an MP4 with the default preset. Excluded: everything above that the mainstream number excludes, plus advanced options, multicam, colour grading beyond presets, interchange and pro formats.

Feature	Weight	Depth	Why
New project, import clips	15%	80%	Drag-and-drop and Import work for phone MP4 / MOV (H.264, HEVC); MKV multi-audio (#601), AVI (#598) fail
Put clips on the timeline, arrange	15%	85%	Drop to timeline (#517), snapping, tracks; new tracks by dropping missing (#483)
Trim, split, delete	15%	80%	Razor, trim, ripple delete; Backspace on gaps (#648) and deselect (#683) confuse
Undo / redo	5%	95%	Snapshot undo for every command
Titles	10%	75%	Type tool, templates; Arabic text broken (#395)
Transitions (cross dissolve)	5%	85%	Drag on, resize (#503)
Music and volume	10%	75%	Clip volume, fades, meters; device issues off macOS
Playback	10%	70%	Smooth on current Macs; lag / CPU reports on 0.5.0 (#523, #525)
Export MP4 with the default preset	10%	80%	Own H.264 + AAC, Quick Export
Save and reopen	5%	85%	.fcproj, auto-save, recovery prompt
Weighted depth	100%	80.0%

Discounts a beginner actually hits:

• Launch and stability ×0.88: the beginner's machine is often a Windows laptop with Intel graphics or an older Mac; open launch crashes there (#380, #413, #477, #512, #582, #655, #661, #687).
• Discoverability and UI clarity ×0.90: a dense Premiere-style interface with no guided start; reports about deselecting, deleting gaps and transitions controls (#683, #648, #577).
• Opening files people send them ×0.95: phone and web video mostly opens; AVI, E-AC-3 audio (#647), FLAC-in-MP4 (#603) and some WebM (#432) do not.

80.0 × 0.88 × 0.90 × 0.95 = 60.2% ≈ ~60%.

User evidence

Counted 2026-10-10 from all 323 GitHub issues and their 270 comments, excluding the four repository collaborators (script in a private scratch folder; regex for praise and for "switched from" phrases, then read by hand):

• Praise: 31 distinct non-maintainer people across 37 issues used praise words. About a dozen are explicit praise, e.g. "Great job making this go!" (#313), "I personally love everything done on it" (#227), "impressive otherwise" (#208), "Great project guys" (#451). Several of them pair the praise with a bug ("great project although it obviously has a lot of bugs", #455).
• Switched from Premiere: 0 reports of having switched. Two people describe testing it as an alternative or hoping to leave Adobe (#411 "testing FilmCraft as an alternative workflow to Adobe Premiere", #325 "finally moving away from adobe would be a god send").
• Open issues: 202 open from non-maintainers; by title, ~93 are bug reports, of which ~64 are core-path (launch, import, playback, timeline, audio, export) and ~29 niche; ~109 are feature requests or questions (Object Mask, plugin compatibility, languages, Android, collaboration, node-based colour).

This supports the discounts: people like the direction and use it, but core-path bugs outnumber niche requests among the bug reports about two to one, and nobody reports having switched yet.

Stage

Alpha. It passes the core-workflow gate (roadmap.md: all six of Premiere's core daily workflows work end to end on macOS and the project saves and reopens). Core workflows run end to end (import, edit, trim, colour, mix, caption, deliver, round-trip through XML / AAF), but ready-for-real-work is ~56%, below the ~75% beta bar, and the standard's beta condition "opens and saves the target's main format reliably" fails outright: FilmCraft cannot open a .prproj (it reads Premiere's FCP7 XML / AAF / OTIO exports instead). Distance to beta: ~19 points of ready-for-real-work plus .prproj import, ~400–700 agent-hours:

To reach beta	Hours
.prproj import of mainstream projects (sequences, clips, effects' common parameters, markers)	60–120
Stability: launch crashes, PR CI running cargo xtask ci on all three OSs, crash triage	40–80
H.264 High 10 / 4:2:2 and HEVC 4:2:2 software decoding (camera media)	30–50
Real-world media corpus and scripted workflow acceptance tests	40–70
UI fidelity top issues (Program monitor handles, Effect Controls, docking, deselect)	60–100
Hardware: zero-copy, VideoToolbox Main 10, Quick Sync / AMF encode, more GPU effects	60–100
Windows / Linux runtime fixes, audio device issues	30–60
Container gaps users hit: E-AC-3, AVI, multi-track audio, still formats	30–50
Interchange validation (AAF / OMF in Avid and Pro Tools, XML speed)	20–40
Performance: lag / CPU reports, AV1 speed	30–60
Total	~400–730

Calibration

Hours are Opus 5.5 agent wall-clock hours, calibrated against this repo's history (git log, gh pr view timestamps):

• Whole history. First commit 2026-09-30; by 2026-10-10, 1,098 commits and 268 merged PRs (243 since 10-05, most from community contributors' agents) took breadth from 0 to ~86%. The 2026-10-01/02 blocks measured ~4–5 agent-hours per checklist point with five parallel agents (old ROADMAP).

• A hardware decode backend (one OS API, H.264 + HEVC, with parity tests): VideoToolbox #33 (first commit 07:25, merged 11:26 the same day), VA-API #358 (+3,942 lines), NVDEC #549 (06:17 → 11:04): ~3–5 h each.

• A hardware encoder per codec and OS: NVENC HEVC #328 (+1,635 lines, 19:28 → 13:46 next day including review), Main 10 HDR #360 (02:38 → 14:05): ~4–8 h each.

• GPU export rendering plus four GPU effects (#421, 15 commits over ~23 h): ~8–12 h.

• A whole-interface language (~3,500 strings, tests): Japanese #449 in one commit: ~4–8 h, plus native review (human). A menus-only language (Ukrainian #407, ~350 strings): ~1 h.

• A clean-room decoder from spec, bit-exact: AV1 (09-30 → 10-01 night, several agents): ~15–30 h.

• Depth work costs more per point than breadth. Each remaining point of ready-for-real-work is a bug class, a format edge or a fidelity comparison rather than a missing menu item; the estimates apply ~1.5–2× the breadth rate.

• Wall-clock limits (from the 2026-10-02 blocks): more than ~5 local agents slows everyone (each worktree build is 6–7 GB; a full cargo xtask ci takes 20–60 min under load; the disk filled twice), and one integrator merging and re-running CI is a bottleneck. With 4–5 parallel agents plus an integrator, wall-clock is ~¼–⅓ of the agent-hours; community PRs now add capacity beyond that.

What parallelizes: almost everything by area (~70%). Serial: .prproj import before beta, fidelity harness before fidelity numbers. Needs a human: the plugin-hosting decision (FFI, licences), native-speaker review of every language, licensed or trained models for AI features, hardware we don't own (Intel Quick Sync, AMD encoders, Blackwell 4:2:2, DeckLink / AJA I/O, control surfaces), and validation in Avid Media Composer / Pro Tools.

Revision history

Date	Change	Summary
2026-10-10	minor	Readiness table: hours to ~95% for each audience (full ~1,450–2,600, mainstream ~500–850, essentials ~180–350); full number confirmed as an additive weighted sum (no change)
2026-10-10	minor	Added the three readiness numbers: full ready-for-real-work formula written out (~55% -> ~56%, same inputs, now computed rather than rounded down), mainstream practitioner ~55% and essentials user ~60% with written weights and discounts, and counted user evidence (31 people praising, 0 switch reports, ~64 core-path vs ~29 niche open bugs)
2026-10-10	minor	Stage checked against the new core-workflow gate (docs/roadmap.md#alpha-gate): passes, stays alpha
2026-10-10	major	Created from ROADMAP.md's scorecard and estimate; full re-measure against Premiere Pro 26.5.2 (menu label diff, installed-bundle listings, 188 open issues); breadth ~87 → ~86% (file formats counted), ready ~50–60% → ~55%, hours re-estimated to include plugins, AI, localization and .prproj; stage alpha
2026-10-05	major	Honest assessment added in ROADMAP.md (checklist ~87% vs ready ~50–60%)
2026-10-02	minor	Scorecard ~62% → ~72% → ~87% as the menu long tail, effects and codecs landed