feat: add on-device summarization helper

Author: prateekjain24

CHANGELOG.md

@@ -5,6 +5,22 @@ All notable changes to ClipDrop will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.6.0] - 2025-09-30
+
+### ✨ Summaries on macOS 26.0+
+
+### Added
+- On-device summarization helper powered by Apple Intelligence foundation models
+- `--summarize` / `-S` CLI flag to append AI summaries to saved text
+- Eligibility checks for content length, format, and platform availability
+- Build automation for `clipdrop-summarize` universal binary and Python bridge tests
+
+### Changed
+- README/usage docs now highlight summarization workflows and requirements
+- Version bumped to 1.6.0
+
+---
+
 ## [1.5.3] - 2025-09-23
 
 ### 🐛 Bug Fixes

README.md

@@ -67,15 +67,25 @@ clipdrop              # → transcript_20240323_143022.srt
 clipdrop meeting.txt  # → meeting notes as plain text
 ```
 
-### 4. 🎥 **YouTube Transcripts**
+### 4. 🤖 **On-Device Summaries** (macOS 26.0+)
+Turn long clipboard text into concise takeaways without leaving your machine:
+```bash
+# Save article and append summary
+clipdrop research-notes.md --summarize
+
+# Skip when content is too short or non-text
+clipdrop notes.txt --summarize
+```
+
+### 5. 🎥 **YouTube Transcripts**
 Research videos efficiently:
 ```bash
 # Copy YouTube URL, then:
 clipdrop -yt                    # Download transcript
 clipdrop -yt lecture.md --lang es  # Spanish transcript
 ```
 
-### 5. 🔒 **Secret Scanner**
+### 6. 🔒 **Secret Scanner**
 Never accidentally save credentials:
 ```bash
 clipdrop config.env -s           # Scan before saving
@@ -127,6 +137,9 @@ clipdrop -yt lecture.md
 
 # Build research document
 clipdrop research.md -a
+
+# Append AI summary (macOS 26.0+)
+clipdrop research.md --summarize
 ```
 </details>
 
@@ -229,4 +242,4 @@ MIT © [Prateek Jain](https://github.com/prateekjain24)
 <p align="center">
   <b>Stop copying and pasting. Start ClipDropping.</b><br>
   <sub>Made with ❤️ for the clipboard warriors</sub>
-</p>
+</p>

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "clipdrop"
-version = "1.5.3"
+version = "1.6.0"
 description = "Save clipboard content (text and images) to files with smart format detection"
 readme = "README.md"
 authors = [

src/clipdrop/__init__.py

@@ -1,3 +1,3 @@
 """ClipDrop - Save clipboard content (text and images) to files with smart format detection."""
 
-__version__ = "1.5.3"
+__version__ = "1.6.0"

src/clipdrop/bin/clipdrop-summarize

@@ -26,9 +26,30 @@ def is_summarizable_content(content: str, detected_format: str) -> tuple[bool, s
         return False, "Content too long for single-pass summarization"
 
     # Heuristic filter to avoid passing obvious code snippets.
-    code_indicators = ("def ", "class ", "function ", "#!/", "import ", "from ")
     lowered = content.lower()
-    if any(indicator in lowered for indicator in code_indicators):
+    code_keywords = (
+        "def ",
+        " class ",
+        "function ",
+        "fn ",
+        "#!/",
+        "import ",
+        "const ",
+        " var ",
+        "public ",
+        "private ",
+    )
+    code_hits = sum(lowered.count(keyword) for keyword in code_keywords)
+    fenced_code = "```" in content or "</code>" in lowered
+    indented_code = any(
+        line.startswith(("    ", "\t")) and any(sym in line for sym in ("(", ")", "=", ":"))
+        for line in content.splitlines()
+    )
+    structural_tokens = any(token in content for token in ("{", "};", "=>", "#include"))
+
+    if fenced_code or code_hits >= 3 or (
+        code_hits >= 2 and (indented_code or structural_tokens)
+    ):
         return False, "Content appears to be code"
 
     if normalized_format in {"md", "markdown", "txt", "text", "html"}:

src/clipdrop/detect.py

@@ -51,6 +51,7 @@ clipdrop log.txt            # Save as text file
 |------|-------------|---------|
 | `--lang CODE` | Language for transcripts | `clipdrop -yt --lang es` |
 | `--chapters` | Include YouTube chapter markers | `clipdrop -yt video.md --chapters` |
+| `--summarize` | `-S` | Append AI summary after saving (macOS 26.0+) | `clipdrop notes.md --summarize` |
 | `--version` | Show version | `clipdrop --version` |
 | `--help` | Show help | `clipdrop --help` |
 
@@ -158,6 +159,19 @@ clipdrop output.txt -f
 # Accept all security prompts
 clipdrop data.json --scan -y
 
+### 9. AI Summaries (macOS 26.0+)
+Generate a quick recap of long clipboard text:
+```bash
+# Save and summarize in one step
+clipdrop article.md --summarize
+
+# Works with existing files too
+clipdrop notes.txt --summarize
+
+# Summaries are skipped for short or non-text content
+clipdrop changelog.txt --summarize
+```
+
 # Combine flags for full automation
 clipdrop log.txt -a -f -y
 ```
@@ -345,4 +359,4 @@ curl api.example.com | pbcopy && clipdrop response.json
 
 ---
 
-For more information, visit: https://github.com/prateekjain24/clipdrop
+For more information, visit: https://github.com/prateekjain24/clipdrop