From c0f4a59b3f6895779e7e124ca9c3dbeab36d0fc8 Mon Sep 17 00:00:00 2001 From: marksverdhei Date: Thu, 19 Mar 2026 11:09:55 +0100 Subject: [PATCH 1/2] docs: list supported file formats in README Closes #14 Co-Authored-By: marksverdhai <249650165+marksverdhai@users.noreply.github.com> --- README.md | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/README.md b/README.md index b50f47c..8713523 100644 --- a/README.md +++ b/README.md @@ -55,6 +55,16 @@ Works with any OpenAI-compatible API: local servers (llama.cpp, Ollama, vLLM, LM - Interactive rename with confirmation - Robust extension handling: isolates name stem from extension for reliable results with smaller models +## Supported Formats + +| Format | How it works | +|--------|-------------| +| **Text files** (.txt, .md, .py, .json, .csv, .xml, .html, etc.) | First 4KB of content sent to the LLM | +| **Images** (JPEG, PNG, SVG) | Base64-encoded and sent via multimodal API | +| **Images** (WebP, BMP, TIFF, GIF) | Converted to PNG via Pillow, then sent as above | + +Audio, video, and other binary files are not supported. Image naming requires a vision-capable model. + ## Requirements - Bash 4+ From 7dc525697e5fad29dfd0347e2e0b06ac5e85db52 Mon Sep 17 00:00:00 2001 From: marksverdhei Date: Fri, 26 Jun 2026 23:26:03 +0200 Subject: [PATCH 2/2] =?UTF-8?q?docs:=20correct=20Supported=20Formats=20?= =?UTF-8?q?=E2=80=94=20audio=20IS=20supported;=204000=20chars=20not=204KB?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Verified against `hat`: - Audio (MP3/FLAC/OGG/AAC/WAV/M4A/Opus/WMA/AIFF/APE) is detected (`is_audio`, magic bytes + extension) and named from `ffprobe`-extracted tags + duration (`collect_metadata` → audio path in `build_user_content`). The previous "Audio … not supported" was incorrect. - Text truncation is `f.read(4000)` (4000 chars), not 4KB. - Note the ffprobe (audio) / vision-model (image) / EXIF (image) requirements. Co-Authored-By: Claude Opus 4.8 --- README.md | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index 8713523..34f0d77 100644 --- a/README.md +++ b/README.md @@ -59,11 +59,14 @@ Works with any OpenAI-compatible API: local servers (llama.cpp, Ollama, vLLM, LM | Format | How it works | |--------|-------------| -| **Text files** (.txt, .md, .py, .json, .csv, .xml, .html, etc.) | First 4KB of content sent to the LLM | -| **Images** (JPEG, PNG, SVG) | Base64-encoded and sent via multimodal API | +| **Text files** (.txt, .md, .py, .json, .csv, .xml, .html, etc.) | First 4000 characters of content sent to the LLM | +| **Images** (JPEG, PNG, SVG) | Base64-encoded and sent via the multimodal API | | **Images** (WebP, BMP, TIFF, GIF) | Converted to PNG via Pillow, then sent as above | +| **Audio** (MP3, FLAC, OGG, AAC, WAV, M4A, Opus, WMA, AIFF, APE) | Not sent to the LLM directly — named from embedded tags (title/artist/album/genre/date + duration) read via `ffprobe` | -Audio, video, and other binary files are not supported. Image naming requires a vision-capable model. +Video and other binary files are not supported. + +Image naming requires a vision-capable model (e.g. `llava`). Audio naming requires `ffprobe` (from FFmpeg) for tag extraction; images additionally include any EXIF metadata in the prompt. ## Requirements