Gentic Data — Documentation

1. Getting Started

Sign Up & Get Your API Key

Before you can use Gentic Data, you need an API key to authenticate your requests.

1. Go to gentic.co/data and create an account.
2. Create an organization from your dashboard. API keys and billing are scoped to the organization.
3. Generate an API key and use it as a Bearer token in your MCP client.

2. Connecting to the MCP Server

The server is available at https://mcp.gentic.co/data. For Claude Code:

claude mcp add gentic-data \
  --transport http \
  https://mcp.gentic.co/data \
  --header "Authorization: Bearer YOUR_API_KEY"

For Claude Web and ChatGPT you can also connect via OAuth — no API key needed. See the connect section on the landing page for other MCP clients (n8n, OpenClaw).

3. Agent Skill

For the best results, pair the MCP server with the Gentic Data agent skill. The MCP server gives your agent tool access; the skill teaches it the optimal workflow order. Both the raw SKILL.md and a ready-to-upload .skill bundle are generated on demand from the live manifest, so they always reflect the current tools and pricing.

Add the skill directly via URL:

https://gentic.co/data/SKILL.md

Or upload a .skill bundle to Claude Managed Agents:

https://gentic.co/data/gentic-data.skill

Download this file and upload it wherever Claude Managed Agents asks for a .skill file. It's a zip bundle generated on demand from the latest SKILL.md.

4. When to Apply

• User wants to create a table or import data from a CSV, Google Sheet, or S3 URL.
• User wants to query, analyze, or explore their data — tables, columns, sample rows, aggregates.
• User wants to preview a table's structure, schema, or row count.
• User wants to update, sync, upsert, or append data from a CSV.
• User wants to insert one or more records into a table with duplicate prevention.
• User wants to search text data semantically or find similar records.
• User asks about their "database", "tables", or "data" in general.

5. Workflow

1. 1. Start with `list_database_tables`

   For any 'what data do I have' question, start with `list_database_tables`. It's free, returns all tables with row counts, and anchors the rest of the conversation. If the user names a table directly, skip to `sample_table` or `get_table_schema`.

2. 2. Import with `create_table_from_csv`

   Accepts HTTPS, S3, Google Sheets, and Google Drive URLs — the last two are auto-converted to direct CSV downloads. Files must be publicly accessible. Table names must be letters/numbers/underscores only. For semantic search, pass `embed_columns` to vectorize text columns — this is the only paid tool at 5¢/row, so mention the cost when importing large datasets. CSV files are capped at 100 MB.

3. 3. Always `sample_table` before writing SQL

   Before running `query_data` or inserting records, call `sample_table` or `get_table_schema` to see the real column names and types. Don't guess. `sample_table` gives you a feel for the data; `get_table_schema` gives you precise types. `query_data` is strictly read-only — only SELECT and WITH (CTEs), no INSERT/UPDATE/DELETE/DROP, no `read_csv_auto()` / `read_parquet()` / `glob()`.

4. 4. Pick the right update tool

   `update_table_from_csv` is the general case with three modes: `replace` (drop + recreate), `append` (add all rows, allows duplicates), `upsert` (add only new rows keyed on `unique_column`). `sync_table_from_csv` is the best answer when the user says "sync", "update", or "refresh" — updates existing rows and inserts new ones in one call. `batch_update_table_from_csv` only touches existing rows (use for corrections). **Always ask which column contains unique identifiers** — don't default to append.

5. 5. Insert records safely

   `insert_record` for a single row, `batch_insert_records` (up to 1000 rows) for many. Both require `unique_column` for duplicate prevention — `insert_record` rejects duplicates, `batch_insert_records` silently skips them. Always `sample_table` first so you know the required columns.

6. 6. Semantic search via `_vectors` tables

   `search_structured` runs cosine-similarity search over a `_vectors` table created with `embed_columns`. Key params: `table_name` must end in `_vectors`, `query` is natural language, `filters` is an optional SQL WHERE clause for structured columns (validated for safety), `limit` defaults to 20 (max 100). Use this when the user asks to find 'similar' items or to search by meaning rather than exact match.

7. 7. Present results clearly

   Don't dump raw JSON. For query results, render a markdown table with column headers; for large result sets, show a head/tail and summarize. For imports and updates, confirm the row counts that changed. For semantic search, lead with the top match and why it's relevant. Always end with a concrete next step (another query, a chained MCP call, an export).

6. Tool Reference

11 tools, rendered live from the Gentic MCP manifest. Parameter tables come directly from each tool's JSON Schema.

7. Pricing

Pricing is pulled live from the Gentic MCP manifest. All prices are per call and deducted from your Gentic credits.

8. Notes

• All tools are organization-scoped — users only see their own database and tables.
• All tools are free **except** vectorized imports via `create_table_from_csv` with `embed_columns` at 5¢/row. A 10k-row vectorized import is $500 — always surface the cost before running.
• Table and column names can only contain letters, numbers, and underscores.
• `query_data` is read-only. Use the insert/update tools for writes.
• CSV files are capped at 100 MB per import.
• Users don't need to 'create a database' — their database is provisioned automatically on first use.