# History Searcher MCP MCP

> Clipboard History Searcher searches through exported clipboard history files (Ditto, CopyQ, Maccy) using your AI client. It quickly finds specific URLs, phone numbers, or code snippets you copied days or weeks ago, making historical data instantly retrievable.

## Overview
- **Category:** productivity
- **Price:** Free
- **Tags:** clipboard-manager, search-utility, data-retrieval, productivity-tool, local-storage

## Description

Ever copy something critical—a unique API key, a long URL, a block of SQL—only to realize you needed it last week? You'll dive into your clipboard manager logs, scrolling through hundreds of entries just to find that one piece of context. This MCP changes that. It takes the exported history file from tools like Ditto or CopyQ and makes every single entry searchable by your AI client. Instead of manually sifting through JSON or plain text exports, you ask your agent what it needs, and it pulls the exact data point, citing where it was found in your past clips. Since Vinkius manages this catalog, connecting to this MCP gives your entire workspace a memory function for everything you've ever copied.

## Tools

### search_clipboard_history
Parses exported clipboard history files (JSON/TXT) to make the contents searchable by your AI client.

## Prompt Examples

**Prompt:** 
```
Find the URL I copied yesterday that had 'github' in it.
```

**Response:** 
```
Found it: 'https://github.com/vinkius/awesome-project' — copied at entry #142.
```

**Prompt:** 
```
Search my clipboard for any phone numbers.
```

**Response:** 
```
Found 3 entries with phone numbers: +1 555-0199, +351 912 345 678, (11) 99999-0000.
```

**Prompt:** 
```
Show me all code snippets in my clipboard history.
```

**Response:** 
```
Found 8 entries that appear to be code: Python (3), JavaScript (4), SQL (1).
```

## Capabilities

### Search historical content
Your AI client parses exported history files (JSON or TXT) and performs targeted searches across all stored clips.

## Use Cases

### Debugging old code
A developer is fixing an issue today but needs the exact API endpoint they copied last week. They ask their agent, 'What was that staging URL I used last Tuesday?' The agent uses `search_clipboard_history` and immediately provides the correct link and its timestamp.

### Writing documentation
A technical writer needs to reference a specific command line argument copied during an early stage of development. They prompt their AI client, 'Show me all code snippets containing '--dry-run' from my history.' The agent pulls the exact syntax needed for the doc page.

### Collecting data points
A student is researching a topic and copies several unique identifiers (ISBNs, product codes) over days. They upload the combined log file and ask their AI client to list them all, getting an organized summary they'd otherwise have to copy manually.

## Benefits

- Stop losing context. Instead of opening your clipboard manager and manually scrolling through logs, just ask your agent to find the data point using `search_clipboard_history`.
- Handles multiple formats. It reads both JSON exports (like from CopyQ) and plain text files (like Ditto), so you don't have to worry about file type compatibility.
- Pinpoints specific data types. Need a phone number or a GitHub URL? Your agent filters the history by pattern, giving you precise results immediately.
- It keeps your context flowing. You can pull important snippets without breaking your workflow, treating your copy history like a searchable database.
- Pure parsing power. This MCP uses native parsing methods; no external dependencies are required to read complex file structures.

## How It Works

The bottom line is you upload the log once; then, you query it repeatedly via natural language prompts.

1. Export your clipboard manager's history into a plain text or JSON file.
2. Connect this MCP to your AI client and provide the exported file for indexing.
3. Ask your agent specific questions, like 'Find all URLs containing '.com'' to get immediate results.

## Frequently Asked Questions

**Which clipboard managers are supported?**
Any manager that exports as JSON or plain text: Ditto (Windows), CopyQ (cross-platform), Maccy (macOS), Clipy (macOS).

**Does it capture images from clipboard?**
No, this parser focuses on textual clipboard entries to save AI context tokens.

**Is my clipboard data sent to the cloud?**
No. Parsed locally. Only text entries are sent to AI context during your session.

**How does `search_clipboard_history` handle different file formats?**
It handles both JSON and plain text files. You can pass in an export from CopyQ (JSON) or a simple TXT dump from Ditto, and the MCP will parse it correctly.

**Does `search_clipboard_history` have limits on how many clips it searches?**
Yes, to prevent context overflow, the tool shows up to 200 clips per search. This limit ensures the data stays within the AI's working memory while still giving you a massive search scope.

**What is required before I can use `search_clipboard_history`?**
You must export your clipboard history file first. The MCP needs a physical input—a JSON or TXT dump—to process the data; it cannot access the active, live clipboard.

**How does `search_clipboard_history` identify code snippets versus regular text?**
The tool uses advanced parsing logic to categorize content. It doesn't just look for keywords; it detects patterns typical of programming languages like Python, JavaScript, and SQL.

**What happens if the clipboard history file I provide is empty or corrupted?**
The MCP handles these situations gracefully. If the input file cannot be parsed or contains no data, it sends a clear error message back to your agent instead of failing.