MCP SERVERKeep a local index of the code and documents you keep re-reading, and search it by meaning instead of by exact wording.
LABSCO SUMMARY
The whole surface is add, list, search, inspect, remove, clear, which is enough for a working set you assemble for one project and throw away afterwards. Updating a document means removing it and adding it again, and anything outside the supported extensions never enters the index at all.
A semantic knowledge base in six tools — documents in, vector search out — over local files, GitHub repositories and direct file URLs.
No account and no key. Documents are capped at 10 MB each, raised through the MAX_FILE_SIZE_MB environment variable.
One command — uvx context-lens
USE CASES
Point the knowledge base at the repository, confirm what actually landed, then ask for a concept rather than a string — the search runs on embeddings, so a question about authentication surfaces the login and credential code even where the word never appears.
Check what the base is holding for a file and when, drop the stale copy, and add the current one back. There is no in-place update, so removal and re-adding is the way a document changes.
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
add_documentAdds a document or GitHub repository to the knowledge base.
Supports local files, GitHub repositories, and direct file URLs.
Supported file types: .py, .txt, .md, .js, .jsx, .ts, .tsx, .mjs, .cjs, .java, .cpp, .c, .h, .hpp, .go, .rs, .rb, .php, .json, .yaml, .yml, .toml, .sh, .bash, .zsh
Maximum file size: 10 MB (configurable via MAX_FILE_SIZE_MB environment variable)
clear_knowledge_baseRemoves all documents from the knowledge base.
Returns: dict: Success status and count of documents removed
get_document_infoGets metadata and statistics about a document in the knowledge base.
list_documentsLists all documents in the knowledge base with pagination support.
remove_documentRemoves a document from the knowledge base.
search_documentsSearches documents using semantic vector similarity to find relevant content.
Uses AI embeddings to understand meaning, not just keywords. Finds related concepts even without exact word matches (e.g., "authentication" finds "login", "credentials").
BEFORE YOU INSTALL
6 tools when we connected.
All 4 documented clients run the server on your machine.
CONTEXT COST
Rated by the size of the 6 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
uvx context-lens{
"mcpServers": {
"context-lens": {
"command": "uvx",
"args": [
"context-lens"
],
"autoApprove": [
"list_documents",
"search_documents"
]
}
}
}{
"mcpServers": {
"context-lens": {
"command": "uvx",
"args": [
"context-lens"
]
}
}
}MCP servers
nicholasglazer
serkan-ozal
caagudelo
roomi-fields