Write your rules in plain English.
“Discounts can’t exceed 20%. Agents can never delete orders.”
SQL · Read-only · Every query traced.

Link any SQL database with a read-only role. Claix reads its structure and drafts a Data Map (tables, descriptions, and relationships) for you to review and edit.




SQL
This is where you set everything up, in two parts:
Create routes.
A route is a retrieval capability: you pick the tables and columns it can use and write a short guide with your rules and vocabulary.
Create your agent.
Give it a name and a description, write its business prompt (what it does and when to use each route), attach its routes, and choose how it answers: raw results or reasoned, typed answers.
No YAML, no frameworks, no code.
Call one endpoint and your agent gets everything it was missing:
Claix Guard
AI agents are moving from reading data to changing it, and today the only protection is a prompt. With Claix Guard, your agent asks before it changes anything, and Claix answers allow or deny, with the reason.
01 Agent proposes
INSERT
INSERT INTO orders (customer_id, status, discount, total) VALUES (4812, 'pending', 0.12, 211.20);
02 Claix checks
03 Decision
Allow
A 12% discount stays within the 20% limit.
“Discounts can’t exceed 20%. Agents can never delete orders.”
Claix turns your words into precise rules, with test cases. You accept, reject or edit them.
Claix checks the statement against your rules and tells the agent why when the answer is no.
The rules are exact.
Claix never touches your data.
The statement, the rule, and the verdict stay on record.
Routes and agents
A route says what can be retrieved: the tables and columns allowed and a guide with your rules. An agent is the identity that uses them: its own endpoint, its own business prompt, and its own routes. Create a route once and attach it to as many agents as you need. A support bot and an internal assistant can share the same database and still see different data.
Agent
restaurant-ops
Route
floor-ops
restaurants · reservations
Route
staffing
employees · shifts
Access controlled
Delegate data access to your agents without handing over your database. Claix connects with a read-only role, and agents only reach the tables and columns you include in a route. Everything outside that perimeter doesn’t exist for them. Queries are validated before they run, and every request is traced so you can review what was asked, which route answered, and what came back.
Guard
Your agent asks before it changes anything. Claix answers allow or deny, with the reason. You write the rules in plain English, and Claix turns them into exact rules. No model decides at check time, and Claix never touches your data.
INSERT
UPDATE
DELETE
BYOK
Use Claix Managed AI and we handle the model, the setup, and the updates. Or connect your own key from OpenAI, Google, Anthropic, or xAI and choose your model. With your own key, the Claix fee drops from €0.10 to €0.02 per question, and you pay your provider’s tokens directly.
With your own key, results depend on the model you choose. Test it in the Sandbox before going live.
Pricing
Add Claix to your system and agents for free, without adding a card
30 successful calls per month
Pay for what you use, no more and no less
30 free successful calls every month
Retrieve
Guard
Scale to infinity and beyond
For teams that need a tailored rollout.
Claix is the controlled data retrieval and security layer for AI agents: it connects to your relational database and lets each agent retrieve only the tables and columns you allow and only execute the actions allowed following your rules and limits, without moving your data. Available via REST API, MCP, and A2A.
Connect any SQL database with a read-only role (Claix drafts a Data Map for you to review), define routes and agents with your rules (no YAML or code), write Guard rules in plain English for the changes an agent may propose, then call one endpoint so your agent gets live typed answers and a verdict before any change, inside the perimeter you set.
No. Claix does not copy your data or build a vector index over it. Agents query your database where it lives, within the tables and columns you allow. Every query is validated before it runs and traced afterward.
No. Claix never writes to your database. Retrieve connects with a read-only role. When an agent proposes an INSERT, UPDATE, or DELETE, Claix Guard answers allow or deny, with the reason, and your code executes. Loading documents into structured fields is a separate product, Claix Ingest, which you choose to use.
Any SQL database, including Postgres, MySQL, and MariaDB, plus Postgres-compatible setups such as Supabase. You connect with a read-only role. Claix never executes writes: Guard only answers whether a proposed change is allowed, and your code runs it.
A route is a retrieval capability: the tables and columns allowed plus a short guide with your rules and vocabulary. An agent is the identity that uses them—its own endpoint, business prompt, attached routes, and response mode (raw results or reasoned typed answers). Create a route once and attach it to as many agents as you need.
POST to /retrieve/{agent_name} with a JSON body of questions (each with an optional format such as string, int, boolean, json, or array:string). You get typed answers in order plus a log_id. The same agent is also reachable over MCP (claix.retrieve.ask) and A2A.
In reasoned mode you can ask for string, int, boolean, timestamp, timestamptz, date, json, array:string, or array:int. Typed values mean your workflow can branch on booleans, use numbers and lists directly, and skip parsing free text.
Every account gets 30 successful calls free each month, with no card. After that, Claix Managed AI is €0.10 per question. With your own LLM key (BYOK) from OpenAI, Google, Anthropic, or xAI, the Claix fee drops to €0.02 per question and you pay your provider’s tokens directly. Results with your own key depend on the model you choose—test in the Sandbox before going live. Claix Guard is €0.01 per call: each check of an INSERT, UPDATE, or DELETE. There is no model and no managed AI on that check.
No. They sit on different axes. “Agentic” is about who triggers retrieval (a fixed pipeline vs an agent calling a tool). “Vector” is about how results are matched (embeddings). An agent can call SQL, BM25, or embeddings; a fixed pipeline can use hybrid search. For relational business data, embeddings are often the wrong default.
Yes. RAG means retrieved information feeds the model. It does not require embeddings, and the retrieved information does not have to be text chunks. A SQL query that returns live records for the model is still retrieval.
When you need semantic discovery over a large permanent text collection, heavy repeated search traffic, or a product where search is the product. For exact values, counts, dates, and live rows already in your SQL database, query the database where it lives with scoped access instead of copying it into an index.
Yes. Claix is not an agent framework—it is a retrieval layer you call from the stack you already use. Call REST from any HTTP client (including n8n), expose the agent as an MCP tool for Cursor and compatible clients, or connect over A2A so other agents can delegate data questions.
The field comes back empty instead of a guess. Queries only run inside the routes you attached to the agent; everything outside that perimeter does not exist for them. Traces show what was asked, which route answered, and what came back.
A generic database MCP connection is quick to wire up, but the access model depends on the server and the role you attach. Claix adds routes, per-agent allowlists, validation before queries run, typed multi-question answers, Claix Guard — which checks every proposed change against exact rules before your code runs — and a reviewable trace, so you define the boundary once and enforce it on every call.
No. MCP defines how an agent calls a tool, not what it is allowed to do. A database MCP server with a broad role gives the agent that role's power.
No. Claix Retrieve is an MCP server (the tool is claix.retrieve.ask), and Claix Guard checks statements headed for any MCP database tool before execute_sql.
When an agent acts in production on data that matters and you want scoped tables and columns, typed answers, and no SQL written by a model. A raw MCP server is still the better choice for development, exploration, and ad hoc analysis.
When your agent proposes INSERT, UPDATE, or DELETE and you want a deterministic yes or no before anything runs. Claix never executes the statement. Your code does, and only after an explicit allow.
No. Structured outputs make the format of the model's answer follow a schema. They do not verify that the values are correct, that the data exists, or that a statement is safe to execute.
Give the agent a guarded function that sends the statement to Claix Guard, returns the reason on deny, and calls execute_sql only on an explicit allow. Do not expose the raw tool. If the check fails, times out, or returns anything other than allow, do not run the statement.
Design the wrapper to fail closed: if the check cannot be completed, the statement does not run.
Connect your database, define the rules, and start retrieving in minutes.