Fabi.ai
Fabi.ai is an AI data analyst platform whose Analyst Agent writes SQL and Python in collaborative Smartbook notebooks, publishes auto refreshing dashboards, and runs Workflows that send scheduled reports and change alerts to Slack, email, Sheets and Slides.
Fabi.ai gives data teams and operators an Analyst Agent that takes a plain English question, writes SQL to pull data from warehouses, databases, files and apps, and writes Python for statistics, modeling and custom charts. The work lands in Smartbooks, AI notebooks with reactive cells and version control, which publish as dashboards and data apps that refresh on a schedule.
Workflows push results to where people work. Scheduled weekly or monthly reports and AI alerts on significant data changes go to Slack or email, and outputs can be written into Google Sheets and Slides. Teams can configure the agent with their business context, set field and table access, and scope agents to approved datasets. The customer chooses OpenAI, Anthropic or a privately hosted model, and an MCP server lets other assistants create analysis threads and save results to Smartbooks. A template gallery offers ready made dashboards and analyses by use case.
Fabi says it is SOC 2 Type 2 certified. Pricing starts with a free Starter plan and a $39 a month Builder seat, with Enterprise on request.
Vendor details
Canonical URL
https://fabi.ai
Category
Data analyst agent
Company status
independent
Use cases & customers
Integrations
Fabi connects to Snowflake, BigQuery, PostgreSQL, Stripe, HubSpot, Salesforce, Google Sheets and other warehouses, databases, files and apps (40+ listed in the template gallery), delivers to Slack, email, Google Sheets and Slides, and offers an MCP server and a Slack app.
In practice
A lean growth team connects its warehouse and customer relationship management data to Fabi, then asks the Analyst Agent plain English questions that return SQL, Python analysis, and a publishable dashboard within minutes.
An operator builds a Smartbook with reactive cells and version control, publishes it as a scheduled dashboard, and lets teammates ask follow up questions through the Fabi agent installed in Slack.
A data lead scopes an analyst agent to only approved tables, connects it through the Model Context Protocol server, and sets real time alerts so stakeholders learn when a key metric shifts.
Agentic Index coverage score
9.0 / 14 capabilities · 64%
| Integrations & Tool Calling | Full |
|---|---|
|
Fabi connects warehouses, databases, files and apps (Snowflake, BigQuery, PostgreSQL, Stripe, HubSpot, Salesforce, Google Sheets and more), writes and runs SQL and Python across merged sources, and its Workflows push results into Google Sheets, Slides, Slack and email through direct connectors. The agent merges several sources in memory, from spreadsheets to data warehouses. The Builder plan includes one data connector and one app connector, and Enterprise includes all of them. SourceFabi.ai, fabi.ai/product/analyst-agent, /product/workflows and /pricingread 2026-10-05 |
|
| Workflow Orchestration | Full |
|
Workflows are flows the customer builds that combine SQL, Python, no code and AI steps, run on schedules and deliver to Sheets, Slides, Slack or email, and Smartbooks chain reactive cells. Fabi does not describe branching or conditions. Smart Reports turn an analysis into a report that refreshes on its own, and dashboard links can be shared with outside viewers who have no paid seat. SourceFabi.ai, fabi.ai/product/workflows and /product/analyst-agentread 2026-10-05 |
|
| Knowledge Grounding & RAG | Partial |
|
The Analyst Agent can be configured with the business's own context and semantics and auto learns as the team engages, and it helps users refine their questions. Fabi does not say how that context is stored or retrieved, or whether it lives anywhere beyond the prompt. SourceFabi.ai, fabi.ai/product/analyst-agentread 2026-10-05 |
|
| Human Oversight & Guardrails | Partial |
|
Admins set field and table access, and scoped agents work only with approved datasets, which bounds what the agent may touch. There is no step where a person approves an analysis or delivery before it goes out. Data teams define exactly which data each agent can reach. SourceFabi.ai, fabi.ai/product/analyst-agentread 2026-10-05 |
|
| Security, Identity & Governance | Full |
|
Fabi says it is SOC 2 Type 2 certified (with GDPR and CCPA readiness badges) and offers granular access control down to fields and tables, with scoped agents limited to approved datasets. Fabi does not describe SSO, the auditor or roles beyond that, and does not publish the SOC 2 report itself. A data processing addendum and a subprocessor list are available, and Enterprise customers can ask for a custom security review. SourceFabi.ai, fabi.ai/security, /pricing and /product/analyst-agentread 2026-10-05 |
|
| Observability & Auditability | Partial |
|
The agent's SQL and Python land as cells in Smartbooks with version control, so the code it ran can be read and traced through reactive cells. There is no audit log or run trace of the agent's steps across users. Smartbooks and Smart Reports keep the analysis and its charts together for review. SourceFabi.ai, fabi.ai, /product/analyst-agent and docs.fabi.ai MCP Serverread 2026-10-05 |
|
| Memory & State Persistence | Partial |
|
Analysis runs in threads backed by Smartbooks that keep their cells and state. The business context the agent learns grounds its answers, and Fabi describes no memory that the agent keeps from one session to the next or how long it would last. SourceFabi.ai, docs.fabi.ai MCP Server and fabi.ai/product/analyst-agentread 2026-10-05 |
|
| Deployment & Data Residency | Not documented |
|
Fabi is a hosted service, and no region selection, VPC, dedicated or self hosted deployment of the platform is described. On Enterprise, Fabi can host a private LLM for the customer or connect to one the customer runs, which changes the model but not where Fabi runs. SourceFabi.ai, fabi.ai/pricing and /securityread 2026-10-05 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
A template gallery offers named templates filtered by use case, industry and data source (marketing pipeline dashboard, deal analysis with AI, closed lost deal analysis, Stripe dashboard, user retention analysis and more), each with its own page. Fabi does not say how a template is copied into a workspace. SourceFabi.ai, Template Galleryread 2026-09-28 |
|
| Triggers & Channel Coverage | Full |
|
Workflows generate scheduled weekly or monthly reports, AI powered alerting detects significant data changes and sends alerts to Slack or email, and dashboards refresh on a schedule, so work starts without a person asking. Custom alerts can also fire on metric thresholds, and people who receive reports by email or Slack need no Fabi login. Scheduled dashboards and workflows start on the Builder plan. SourceFabi.ai, fabi.ai/product/workflows and /pricingread 2026-10-05 |
|
| Model Flexibility & Routing | Full |
|
Teams run analysis on the engine they prefer, OpenAI, Anthropic or their own privately hosted LLM (on Enterprise), and the MCP server plugs the analyst into the LLM of the customer's choice. On Enterprise, Fabi can host the private LLM itself or connect to one the customer runs, and the Builder plan includes the latest models. SourceFabi.ai, fabi.ai/product/analyst-agent and /pricingread 2026-10-05 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
The MCP server at app.fabi.ai/mcp (bearer token or OAuth) exposes tools to create an analysis thread, submit a request, poll for the result and save the generated cells to a Smartbook, so outside assistants work with Fabi's own objects. Fabi calls it a plug and play server for connecting the analyst to an LLM platform. A Slack app also lets people ask questions of their data inside Slack. SourceFabi.ai, docs.fabi.ai MCP Server and fabi.ai/product/analyst-agentread 2026-10-05 |
|
| Testing, Debugging & Optimization | Not documented |
|
Fabi publishes no test set, scored cases or evaluation of the agent's answers. Users can read and edit the agent's SQL and Python in Smartbooks, which helps them debug an answer, and an AI assistant suggests code while they wrangle data. SourceFabi.ai, fabi.ai/product/analyst-agent and /product/workflowsread 2026-10-05 |
|
| Browser & Computer Use | Not documented |
|
There is no browser, desktop or computer control. Fabi works through data connectors, SQL and Python. It writes code against connected sources and sends results through direct connectors to Sheets, Slides, Slack and email. SourceFabi.ai, fabi.ai and /product/workflowsread 2026-10-05 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Free Starter plan; Builder $39 per month
per builder seat per month with an AI request allowance
What is public
Fabi publishes the Starter and Builder prices and limits; Enterprise is on request.
Variable cost rationale
Builder is a flat monthly seat with a request cap rather than metered overage; heavier use moves a team to Enterprise.
Overage / add-ons
AI requests capped by plan (500 a month on Builder); Enterprise unlimited
Sales call required
Mixed (some tiers require a call)
Free / trial
Free Starter plan (10 AI requests a day)
Lowest paid plan
Builder ($39 per month)
Related vendors
- Actian AI Analyst — Wobby, now sold as Actian AI Analyst, answers plain language…
- AnswerRocket — Enterprise AI agent platform for analytics built around Max, which…
- AskEarth — Geospatial intelligence agent (GIA) that answers plain language…
- AskYourDatabase — AskYourDatabase is a conversational data analyst that turns plain…
- Athenic AI — Athenic AI is a data analyst agent that connects databases,…
- BlazeSQL — BlazeSQL is an AI data analyst that turns plain English into SQL…
Alternatives to Fabi.ai
The closest documented capability profiles to Fabi.ai among data analyst agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Chord8.0 / 14Fuller documented coverage on Knowledge Grounding & RAG and Human Oversight & Guardrails
- ThoughtSpot12.0 / 14Adds documented Deployment & Data Residency and Testing, Debugging & Optimization
- Zenlytic11.0 / 14Adds documented Deployment & Data Residency and Testing, Debugging & Optimization
- AskYourDatabase7.5 / 14Adds documented Deployment & Data Residency
- Datafold11.5 / 14Adds documented Deployment & Data Residency and Testing, Debugging & Optimization
- Datapad7.5 / 14Adds documented Deployment & Data Residency
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded