[Feature] Hashtag tag parsing in /capture and remember

[rahilp]

Hashtag tag parsing in /capture and remember

Why

Tags currently require a JSON array. That’s fine for API calls but it’s a real barrier for everyday capture — nobody types {"tags": ["health", "fitness"]} on their phone.

Hashtags fix this. "Went for a run today #health #fitness" should just work.

What this covers

• Parse hashtags from content in the /capture REST endpoint
• Parse hashtags from content in the remember MCP tool
• Strip hashtags from stored content so the text stays clean
• Merge extracted hashtags with any explicitly provided tags (deduped, case-insensitive)

Implementation

function extractHashtags(content: string): { cleanContent: string; hashtags: string[] } {
  const hashtags = (content.match(/#(\w+)/g) || []).map(t => t.slice(1).toLowerCase());
  const cleanContent = content.replace(/#\w+/g, '').trim();
  return { cleanContent, hashtags };
}

Apply in both /capture and remember before duplicate check and storage:

const { cleanContent, hashtags } = extractHashtags(c);
const finalTags = [...new Set([...explicitTags, ...hashtags])];

Acceptance criteria

• "note content #tag1 #tag2" stores with tags ["tag1", "tag2"] and clean content
• Explicit tags in JSON body merge with hashtag-extracted tags
• Duplicate tags deduped (case-insensitive)
• Works in both /capture REST endpoint and remember MCP tool
• No breaking change — existing calls without hashtags behave identically

Why this matters

Enables natural capture from iOS Shortcuts, browser bookmarklet, and voice input without any JSON formatting. Critical for personal use cases — health logging, relationship notes, travel — where users capture quickly on mobile.