The agent-native knowledge base
Give your AI agent long-term memory. Ingest documents, web pages, emails, Slack messages, and raw text — then search everything with natural language. Powered by vector embeddings via Gemini.
How it works
From sign-up to semantic search in seven steps.
Sign up and get an API key
Create a free account at gentic.co, generate an API key from your dashboard, and use it as a Bearer token to authenticate your agent.
Connect your agent
Add the MCP server URL to Claude Code, Claude Web, OpenClaw, Claude Cowork, ChatGPT, n8n, or any MCP-compatible client. Pass your API key as a Bearer token.
Ingest your content
Vectorize documents (PDF, DOCX, TXT), web pages, emails, Slack messages, reviews, or any raw text. Content is chunked and embedded automatically.
Build your knowledge base
Keep feeding content over time — product docs, customer feedback, creator notes, campaign briefs. Each piece is chunked (~1,000 chars) and stored with source metadata.
Search with natural language
Ask your agent anything and it searches your entire knowledge base semantically. Filter by source type, category, or document. Results are ranked by relevance.
Manage your sources
List all indexed sources to see what's in your knowledge base. Delete content by document or specific chunks when it's no longer relevant.
Build agent workflows
Chain knowledge tools with other Gentic MCP servers — enrich influencer outreach with brand docs, ground ad creative briefs in past campaign learnings, or power any RAG pipeline.
Available tools
Six tools your AI agent can call through the Model Context Protocol. Three for ingestion, three for retrieval and management.
vectorize_document
Vectorize a document file — PDF, DOCX, TXT, or RTF. Accepts HTTP URLs and Google Drive links. Content is chunked, embedded, and indexed asynchronously.
- -PDF, DOCX, TXT, RTF support
- -Google Drive link support
- -Async processing via workers
vectorize_content
Vectorize raw text content — emails, Slack messages, reviews, social posts, creator feedback, or notes. Each chunk is ~1,000 characters.
- -Email, Slack, review, social, notes
- -Custom source type & category
- -Automatic chunking & embedding
vectorize_web_content
Scrape a web page, convert to clean text, chunk, and embed. Includes a $0.05 scraping fee on top of per-chunk costs.
- -URL scraping & text extraction
- -Clean HTML-to-text conversion
- -Automatic chunking & embedding
search_knowledge
Semantic search across all indexed content using natural language. Filter by source type, category, or document. Results ranked by cosine similarity.
- -Natural language queries
- -Filter by source type & category
- -Up to 50 results per query
list_kb_sources
List all indexed sources showing document ID, title, source type, chunk count, category, and creation date. Get a bird's-eye view of your knowledge base.
- -Document titles & source types
- -Chunk counts per document
- -Category & date metadata
delete_content
Soft-delete content by document ID or specific chunk IDs. Content is excluded from search results but retained in the database.
- -Delete by document or chunk
- -Soft-delete (recoverable)
- -Immediate search exclusion
Connect in seconds
Sign up, grab your API key, and add the server to your agent. Ingest content, search your knowledge base, and manage sources. Works with Claude Code, Claude Web, OpenClaw, Claude Cowork, ChatGPT, n8n, and any MCP client.
claude mcp add gentic-knowledge \
--transport http \
https://mcp.gentic.co/knowledge \
--header "Authorization: Bearer YOUR_API_KEY"You can also connect via OAuth — just add Gentic Knowledge as a connector in Claude or ChatGPT settings. No API key needed; authentication is handled automatically.
Install the Gentic agent skill to teach your agent the optimal workflow — ingesting content, searching knowledge, managing sources — so it gets the best results automatically.
npx skills add gentic-co/agent-skillsWorks with Claude Code, Cursor, Copilot, and 40+ other agents.
Pay per chunk
Ingestion is billed per chunk (~1,000 characters each). Search and management tools are free. No subscriptions.
| Action | Cost |
|---|---|
| vectorize_document | $0.05 / chunk |
| vectorize_content | $0.05 / chunk |
| vectorize_web_content | $0.05 / chunk + $0.05 scraping |
| search_knowledge | Free |
| list_kb_sources | Free |
| delete_content | Free |
Example cost
A 10-page PDF produces roughly 30 chunks. At $0.05 per chunk, that's $1.50 to ingest. Once indexed, every search query against it is free — no matter how many times your agent searches.
Why agent-native?
Traditional knowledge bases weren't built for AI agents. This one was.
No UI to learn
Your AI agent is the interface. Say "index this PDF" or "search for our refund policy" in natural language — no dashboards or vector DB consoles required.
Composable workflows
Chain knowledge tools with influencer search, creative generation, data queries, and any other MCP server. Ground every agent action in your own content.
Portable
Works with any MCP-compatible client — Claude Web, Claude Code, OpenClaw, ChatGPT, n8n, and more. Switch clients without changing a thing.
Pay per chunk
Only pay when you ingest content. Search is always free. No monthly seats, storage fees, or platform subscriptions.