Rewrite for direct use with different LLM providersHEADmaster

1 files changed, 45 insertions, 0 deletions

diff --git a/README.md b/README.md
new file mode 100644
index 0000000..36d5ffc
--- /dev/null
+++ b/README.md
@@ -0,0 +1,45 @@
+# tldw
+
+Too Long; Didn't Watch - Summarize YouTube videos using Claude.
+
+## Installation
+
+```bash
+uv sync
+```
+
+## Usage
+
+### tldw - Summarize a video
+
+```bash
+tldw <youtube-url>
+tldw -l es <youtube-url>                            # Use Spanish captions
+tldw -r <youtube-url>                               # Output raw transcript
+tldw -m openai/gpt-4o <youtube-url>                 # Use OpenAI
+tldw -m anthropic/claude-sonnet-4-20250514 <url>    # Use Anthropic API
+```
+
+**Options:**
+- `-l, --language` - Language code for captions (default: `en`)
+- `-r, --raw` - Output raw transcript without summarizing
+- `-m, --model` - Model in `provider/model` format (default: `claude/sonnet`)
+- `-s, --system-prompt` - Custom prompt for the LLM
+
+**Supported providers:**
+- `claude/*` - Claude Agent SDK (sonnet, opus, haiku)
+- `anthropic/*` - Anthropic API (requires `ANTHROPIC_API_KEY`)
+- `openai/*` - OpenAI API (requires `OPENAI_API_KEY`)
+
+### yt-transcript - Get raw transcript
+
+```bash
+yt-transcript <youtube-url>
+```
+
+Outputs timestamped transcript snippets.
+
+## Requirements
+
+- Python 3.12+
+- Claude Code CLI (bundled with multillm-claude)