AI for Database

What AI for Database is

AI for Database is a natural-language analytics tool that sits on top of a database or spreadsheet. Its headline reads "Ask my database anything. In plain English," and the company's own description is "Instant answers from your live database or spreadsheet, plus one-click dashboards and alerts when data changes. No SQL, ever." The home page labels it "For teams without a data team," and the use-case menu names sales, operations, product managers and founders. The company's blog adds customer success to that list and says the tool suits teams where most users don't write SQL.

So the builder here isn't a canvas or a form designer. You build 3 things by talking to your data: answers, dashboards and automated checks. The footer says it is a product of Vynara (formerly Wavicle Technologies), which lists several other products of its own.

What you build and how

Connecting. The databases menu lists PostgreSQL, MySQL, MongoDB, SQL Server and Google Sheets, with a "view all" link for more, and a stack strip on the home page adds SQLite, Oracle, Redis, Firebase, Slack, Salesforce, HubSpot, Notion, Airtable, AWS, Azure and GCP. Actions go out to Slack, email, webhooks, Google Sheets and CRMs, and deployment options are AWS, Azure, GCP or self-hosted. The free start needs no credit card and nothing to connect, because sample data is included, and connections are read-only by default. The flow on the page's closing banner is "Connect, Ask, Pin, Automate."

Asking. You type a question in plain English and the tool answers from live data. One example on the site asks about last week and requests a breakdown by source; the tool, it says, "runs a series of queries and walks you through what changed." Answers come with charts, follow-up questions are supported, and results export to Sheets or CSV. Each answer in the home page demo carries its receipt: the tables touched, rows scanned, run time and a View SQL link. The Pro plan lists "Save queries and inspect the SQL," so a person who does read SQL can check what the AI wrote. The site's tagline is "The assistant that shows its work."

Dashboards and reports. The Pro plan includes exploring answers "with charts & dashboards." The workflow is to pin a good answer to a dashboard. The company's blog says its dashboards are self-refreshing and built from plain-English questions. The Team plan adds sharing dashboards and workflows with colleagues.

Workflows and alerts. This is the automation layer, listed in the menu as database workflows and alerts and monitoring. The example on the site is "Email me every Monday if any support ticket has been open for more than 7 days." The tool "creates a scheduled workflow that checks your data and emails or calls a webhook when the condition is met." The Pro plan lists scheduled workflows and webhook triggers. A workflow reads like a sentence: the home page example emails operations with the reorder list when inventory for any SKU drops below 10 units, checked hourly, with a dry run first and every fire logged. Actions listed for workflows are Slack alerts, email notifications, webhook triggers and CRM updates.

Working from an AI assistant. The home page also describes asking from inside your own assistant, with examples such as listing your dashboards and how many AI credits you have left. It states that "Queries are read-only. Workflows can send email or webhook alerts, so you can set those tools to ask for approval in your assistant." The page names Claude, ChatGPT and Muse by Meta as assistants you can connect through its MCP server, and lists what you can do from there: ask questions in plain English, have the tool read and explain an unfamiliar schema, run your own SQL inside a read-only transaction, investigate a metric, set up alerts, and check dashboards and remaining AI credits.

Public API for agents. A public REST API uses scoped API keys and an OpenAPI 3.1 contract. The page sets out a 6-step path for an agent: discover connections and cached schemas, validate with read-only SQL, draft a paused workflow, preview it without contacting external systems, test one live delivery, then publish an immutable version and inspect its run history.

Permissions and control. The FAQ on the home page covers what the AI can read versus write: "You control permissions. Start with read-only, then enable specific actions as needed." On approvals: "You set the rules. Require approval for high-risk actions, allow auto-execution for safe workflows." The site says every query, dashboard and action is logged, and audit logs can be exported for compliance. Credentials are "encrypted at rest with AES-256" and "never stored in plain text," and you can rotate keys anytime. The trust section shows badges for AES-256 encryption, "SOC 2 ready," "GDPR ready" and audit logging, and says your data never trains anyone's model.

Hosting. The FAQ gives you a choice. Self-hosted "keeps everything in your infrastructure." Cloud-hosted "encrypts data in transit and at rest." On self-hosting the site says "Yes. Full self-hosted deployment available. One-time setup fee plus annual license." A separate banner reads "Prefer to self-host? Run AI for Database on your own infrastructure."

What end users get. For the people who use the finished result, the output is a pinned dashboard or an alert in their inbox. On the Team plan you can assign connections to team members, so each person works against the data they should see.

Plans, pricing, and what the company claims

Pricing is credit based. The home page shows monthly and yearly toggles, with yearly billing saving 20%.

• Free. The page says "Free plan available" and promises "No credit card. Free models forever" with a "Try it free" path and sample data included. The blog says you can start with a free account and connect an existing database with no payment to test the core functionality.
• Pro costs $20 per user per month, billed monthly, and includes $20 in credits per user per payment. The credits are yours to use and don't expire. It includes asking your databases in plain English, charts and dashboards, saved queries with the SQL visible, scheduled workflows, webhook triggers and colleague invites.
• Team costs $50 per user per month, billed monthly, with $50 in credits per user per payment. It lists "2.5× the credit funding per user," an individual balance for each member, assigned connections, usage tracking by member, audit-log review, and shared dashboards and workflows.
• Enterprise is custom: custom usage and seat requirements, review of company SSO requirements, security and compliance discussions, deployment planning, support terms and procurement and contract support.

Credits "never expire," with no weekly limits or monthly resets, and unused credit stays in your personal balance. You can top up $10, $25, $50 or $100 at any time. When the balance runs out, "AI usage pauses." Prices and credits are per user and exclude applicable taxes. The company also offers a demo.

Who it suits and who it does not

This tool fits a founder, a customer success lead or an operations team that has a database and no analyst. The blog places it with a small team of 10-50 people and a budget of $0-$500 a month, starting on the free tier and moving to the paid tier once you need automated alerts. The credit model means cost follows usage per person rather than a flat platform fee, which is easy to forecast for a small team.

It also fits teams that want data kept on their own infrastructure, because the self-hosted option does that, though that route involves a setup fee and an annual license.

Be clear about the limits. The company's own blog says it is the wrong pick if what you mainly want is a row editor for manual data fixes, and it describes the tool as read-oriented analytics. It is not a general app builder or form tool. Dashboards are built from questions rather than designed on a canvas. Because AI usage draws on a credit balance, heavy users on a plan will want to watch usage and top up.