UI Bakery
Also known as: UI Bakery, uibakery.io, UI Bakery AI App Agent, UI Bakery Agent, UI Bakery AI
Internal tools platform whose AI App Agent turns a prompt into a working React app on the customer's data, with per tool ask or allow controls over migrations, data actions and screenshots, Git, checkpoints and self hosting.
UI Bakery is a platform for internal software that offers two ways to build: a drag and drop low code editor and an AI only mode driven by its AI App Agent. In the AI mode a builder describes an app in plain text or attaches a screenshot of something similar, and the agent generates a working React application, connects it to the customer's data sources or a hosted PostgreSQL database, and keeps iterating through chat. The builder can open the code directly, revert to any checkpoint, connect a Git repository and release the result to chosen environments.
The agent works with a set of tools the customer controls one by one. It can apply database migrations when requirements change, take screenshots of the rendered app to see what users see, execute actions against connected data sources to inspect real data, and sync a data source's structure when columns or tables change. Each tool can be set to always allow or always ask, and on ask the agent pauses in the chat for an allow once, always allow or deny decision. Running actions against data defaults to ask. An AGENTS.md file holds app level instructions the agent applies on every build, and the agent reads app logs to fix the issues it finds.
Generated apps can use built in UI Bakery authentication, which brings roles, SSO and audit logs, or a custom login. The platform runs on UI Bakery's cloud, on a dedicated instance in a region the customer chooses, or self hosted on Docker or Kubernetes, including air gapped installs.
The runtime AI inside the apps it builds answers and does not act. AI actions generate, summarize, classify and extract text. Self hosted installations still draw AI credits from a linked cloud workspace unless an Enterprise customer brings its own model provider key.
Vendor details
Canonical URL
https://uibakery.io
Category
Coding agent
Subcategory
AI app builder for internal tools
Funding status
Private (UI Bakery, Inc.).
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Generated apps connect to 45+ data sources, including PostgreSQL, MySQL, MongoDB, Snowflake, BigQuery, Salesforce, HubSpot, Stripe, Slack, GitHub, OpenAPI and HTTP APIs, plus a hosted PostgreSQL database the agent can migrate. The agent connects to Git for source control and can run actions against connected data sources under per tool permission settings. An MCP server (beta) lets external AI agents build apps and manage the workspace.
In practice
An operations lead prompts the agent for an employee portal on the company's PostgreSQL database, reviews each proposed migration before it runs, and releases the app to production.
A developer keeps app level coding and design rules in AGENTS.md so every generation follows them, then edits the React code directly where needed.
A regulated team runs UI Bakery self hosted and air gapped, bringing its own model provider key on the Enterprise license.
Sources & related URLs
Related / legacy domains
Agentic Index coverage score
10.0 / 14 capabilities · 71%
| Integrations & Tool Calling | Full |
|---|---|
|
The agent takes authenticated action in outside systems through its tools. It applies migrations to the app database, executes actions against connected data sources (query a database, check system state, call an API), syncs data source structure, and connects generated apps to 45+ data sources and a Git repository. The executing actions tool is scoped to inspection for answering requests. Sourcedocs.uibakery.io/build-with-ai/features.mdread 2026-10-02 |
|
| Workflow Orchestration | Partial |
|
A single agent runs multi step execution. It shows each step it takes while generating an app, reads logs to address issues and applies migrations along the way. There is no multi agent execution, no subagents and no orchestration of the agent inside workflows. Platform automations schedule app actions, not the agent. Sourcedocs.uibakery.io/build-from-scratch/custom-appread 2026-10-02 |
|
| Knowledge Grounding & RAG | Partial |
|
The agent assembles context from the app itself. It knows the app's schema, code and configuration, keeps a synced model of each data source's structure, and can inspect real data through actions. The agent has no maintained retrieval structure over the customer's documents or knowledge (index, embeddings layer or knowledge base). Sourcedocs.uibakery.io/build-with-ai/features.mdread 2026-10-02 |
|
| Human Oversight & Guardrails | Full |
|
Each agent tool (migrations, screenshots, executing actions, syncing data sources) can be set per app to Always ask, so approval comes before an action commits. On ask, the agent pauses in the chat for Execute or Reject on migrations, and for Always allow, Allow once or Deny once on the others. Executing actions against data defaults to Always ask. Checkpoints revert the app, and releases are an explicit step. Sourcedocs.uibakery.io/build-with-ai/features.mdread 2026-10-02 |
|
| Security, Identity & Governance | Full |
|
Builder has built in roles, Team adds role based access control and audit logs, and Enterprise adds SSO through Okta, Azure AD, SAML, OpenID and OAuth 2.0 with role mapping and SCIM 2.0. Generated apps can inherit workspace SSO, RBAC and audit logs. The homepage states SOC 2 compliance with a Type 2 badge, and the docs link a Drata trust portal alongside security and AI security pages. Sourceuibakery.io, uibakery.io/pricing and docs.uibakery.ioread 2026-10-25 |
|
| Observability & Auditability | Full |
|
The builder shows each step the agent takes while generating an app, and tool permission prompts appear in the chat history. A Logs panel shows current app logs, checkpoints record each prompt's resulting state, release history lists every released version, and audit logs record user actions on Team and above. UI Bakery does not say how long agent step history is kept. Sourcedocs.uibakery.io/build-from-scratch/custom-appread 2026-10-02 |
|
| Memory & State Persistence | Partial |
|
The agent keeps conversation state within a chat, with checkpoints to revert the app to an earlier state and a New chat option to start fresh. AGENTS.md holds app level instructions the agent applies on every build, which are instructions, not memory. There is no memory layer with a set scope and lifetime. Sourcedocs.uibakery.io/build-with-ai/agentread 2026-10-02 |
|
| Deployment & Data Residency | Full |
|
Enterprise runs a dedicated execution environment and database in the region of the customer's choice, and the platform self hosts on Docker, Kubernetes and cloud VMs, including air gapped and multi instance deployments. Self hosted installations still draw AI credits from a linked cloud workspace unless an Enterprise customer brings its own model provider key, and the Custom App feature on premises requires contacting support. Sourceuibakery.io/pricingread 2026-10-02 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
Builders get customer adoptable starting points for the agent's work. The Build with AI flow lets a builder select from ready made templates instead of writing a prompt, and app recipes (pre built templated apps) cover internal tools, dashboards and portals. Sourcedocs.uibakery.io/build-with-airead 2026-10-02 |
|
| Triggers & Channel Coverage | Partial |
|
Work reaches the agent only when someone starts it, through prompts in the builder chat, and external AI agents can drive the workspace through the MCP server (beta). Nothing starts the agent on its own from an event, schedule or webhook. Scheduled jobs and webhooks run app actions, not the agent. Sourcedocs.uibakery.io/build-with-ai/agentread 2026-10-02 |
|
| Model Flexibility & Routing | Partial |
|
In the cloud the agent offers a Fast and a Smart setting without naming the model makers. UI Bakery's AI security page says its AI is based on OpenAI GPT and Claude models, so UI Bakery decides the routing across makers. Self hosted admins can supply their own Anthropic key and set the agent's models, all from that one maker. The replaceable AI datasource applies to AI actions inside generated apps, not to the agent. Sourcedocs.uibakery.io build-with-ai/agent, reference/security/ai-security and on-premise self-hosted-ai-chat-setupread 2026-10-25 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
An Instance API automates app releases and a SCIM 2.0 API manages user accounts on self hosted installations. Webhooks let other services start app actions, and an MCP server (beta) lets external AI agents build apps and manage the workspace. Sourcedocs.uibakery.io/llms.txtread 2026-10-02 |
|
| Testing, Debugging & Optimization | Partial |
|
The agent checks its own work. It reads app logs to fix the issues it finds and uses the screenshot tool to compare rendered UI against intent, and builders can revert to checkpoints. There is no evaluation harness, scored test cases or quality gate in the release path. Sourcedocs.uibakery.io/build-with-ai/features.mdread 2026-10-02 |
|
| Browser & Computer Use | Not documented |
|
The agent does not control a browser, desktop or remote computer. The screenshot tool comes closest, letting the agent inspect the rendered state of the app it is building, but that is viewing its own output, not operating an interface. Sourcedocs.uibakery.io/build-with-ai/features.mdread 2026-10-02 |
|
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
Product updates
Outside AI agents connected to the UI Bakery MCP server can now reach audit logs, usage data, browser sessions, workspace settings and data source synchronization, in cloud release 3.205.0. The same release lets administrators view and revoke the active sessions of workspace members.
Bears on: MCP / tool calling / API
View sourceCloud release 3.202.1 lets vibe apps, UI Bakery's AI generated app type, be updated through the UI Bakery MCP server and through Git. Outside coding agents connected to the MCP server can now edit these apps, and their changes can be kept in Git source control.
Bears on: Integrations
View sourcePricing
Builder at $20 per developer per month billed yearly ($25 monthly). A free plan is available.
Per developer seat, each carrying a monthly allowance of AI usage credits. Viewer seats are capped by tier, and automation executions beyond the plan allowance are sold in paid blocks.
Included quota
Builder includes $25 of AI usage credits per developer a month, up to 50 workspace viewer seats and 5,000 automation executions. Team includes $40 of credits per developer, RBAC and audit logs. Free includes trial AI credits and 1,000 automation executions.
What is public
Cloud and self hosted plan tables, per developer prices, AI credit allowances, viewer seat caps, automation limits and full feature comparisons are published. Enterprise is custom.
Billing mechanics
Per developer seat, billed yearly (20 percent off on Builder, 12.5 percent off on Team) or monthly, with a monthly AI credit allowance per developer. Public app users are unlimited on every plan.
Cost watchouts
AI credits per developer are the unit of agent work, and heavy generation can exhaust them. How extra credits are bought is not published. Self hosted AI use is billed through a linked cloud workspace. Automations beyond the included executions cost $50 per 5,000. SSO, SCIM and more than 50 viewer seats require Enterprise.
Variable cost rationale
Seats are a fixed monthly price per developer. What varies is agent work, which draws down each developer's monthly AI credit allowance, and automation runs, which past the plan allowance cost $50 per extra 5,000 executions. No top up price for AI credits is published, so nothing public says what heavy generation costs once the allowance is spent. Public app users are unlimited on every plan, so a growing audience does not raise the bill, while more than 50 viewer seats moves a team to Enterprise.
Additional watchouts
Seat prices cover developers only. Agent heavy work runs on the AI credit allowance, and the price of credits beyond it is not published.
Overage / add-ons
Automation executions beyond the plan allowance cost $50 per extra 5,000. An overage rule for AI credits is not published.
Sales call required
Mixed (some tiers require a call)
Free / trial
Free plan with unlimited apps and monthly AI trial credits, cloud or self hosted
Lowest paid plan
Builder at $20 per developer per month billed yearly ($25 monthly)
Commercial notes
Self serve signup on Free, Builder and Team; Enterprise through sales, which also sells app migration and custom development services.
Key ambiguities
AI credits are used for development with the AI App Agent, but neither the price of extra credits nor what one credit buys is published.
Support SLA / resale
Free includes community support. App migration and custom development services are sold through sales alongside Enterprise.
Missing data
Extra AI credit price and the Enterprise floor.
Related vendors
- Cognition — Maker of Devin, an autonomous AI software engineer
- 10Web — Agentic website platform whose AI agents build, host and manage…
- AgentUI — Managed AI app builder for operations teams that generates and hosts…
- Aider — Open source, model agnostic terminal coding agent that edits your…
- Amp — Amp is a coding agent from Amp Frontier Corporation that runs each…
- Anthropic Claude Code — Anthropic's agentic coding system across terminal, desktop, IDE, web…
Alternatives to UI Bakery
The closest documented capability profiles to UI Bakery among coding agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Moderne10.5 / 14Fuller documented coverage on Knowledge Grounding & RAG
- Bubble10.0 / 14Fuller documented coverage on Workflow Orchestration
- Zite9.5 / 14Fuller documented coverage on Model Flexibility & Routing
- 10Web9.0 / 14Fuller documented coverage on Workflow Orchestration
- Augment Code12.0 / 14Fuller documented coverage on Workflow Orchestration and Knowledge Grounding & RAG
- Charm10.0 / 14Fuller documented coverage on Triggers & Channel Coverage and Model Flexibility & Routing
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded