How much does Bito cost?

AI code review on Team is $12 per seat a month billed annually or $15 monthly, for up to 25 seats with 5,000 lines per seat a month and $5 per extra 1,000 lines. Professional is $20 per seat a month billed annually or $25 monthly, with unlimited seats and a 14 day free trial needing no card; self hosting is a $5 per seat add on. Enterprise is custom. AI Architect and Governor are priced through sales on the size of the indexed codebase, with free trials.

Does Bito support MCP and tool calling?

Full under the Agentic Index verification standard. Bito publishes public evidence of Model Context Protocol (MCP) support, tool calling, and API or SDK extensibility for connecting agents to external tools. Evidence basis: Governor exposes OpenAI and Anthropic compatible endpoints for messages, chat completions and responses at gateway.bito.ai, with an admin command line tool for managing routers and keys. AI Architect runs an MCP server for each workspace th. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does Bito have SOC 2, SSO and role based access control?

Full under the Agentic Index verification standard. Bito publishes public evidence of enterprise security and identity controls, covering a compliance posture such as SOC 2 or ISO 27001, single sign on (SSO or SAML), and role based access control (RBAC). Evidence basis: Workspaces have Owner, Admin and Member roles, and Enterprise adds single sign on, with AI Architect supporting SAML 2.0 and OIDC through Descope and adding new users on first sign in. Bito holds SOC 2 Type II, with the report shared on req. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Can Bito be self hosted or run in a private VPC?

Full under the Agentic Index verification standard. Bito publishes public evidence of self hosted or private VPC deployment, along with data residency options for controlling where agent data is processed and stored. Evidence basis: The hosted service runs in the United States. Governor can be self hosted on Docker or on Kubernetes with Helm, AI Architect runs self hosted in standalone or enterprise mode, code review self hosting is a $5 per seat add on for Professiona. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

Bito

Q: Who is Bito for?

engineering organizations running several coding agents, platform and developer experience teams managing model spend, teams wanting automated pull request review on GitHub, GitLab or Bitbucket, companies that need coding agent tooling self hosted

Also known as: Bito AI, Bito AI Architect, Bito Governor, Bito AI Code Review Agent, bito.ai

Visit site

Agent infrastructureindependentVerified 2026-11-08

Entry priceAI code review on Team is $12 per seat a month billed annually, or $15 monthly. Professional is $20 per seat billed annually, or $25 monthly. Enterprise is custom, and AI Architect and Governor are quoted by sales.Full pricing detail

Bito sells AI tools that work alongside a team's coding agents. AI Architect builds a knowledge graph of the codebase, tickets and docs, serves it to Claude Code, Cursor and Copilot over MCP, and turns approved Jira plans into branches and pull requests. The AI Code Review Agent reviews every pull request on GitHub, GitLab and Bitbucket, and Governor routes each coding agent request to a right sized model on the customer's own keys.

Bito, founded in 2021 by Amar Goel, Anand Das and Mukesh Agarwal and sold by Bito Inc. of Menlo Park, California, makes AI tools for software teams that sit beside the coding agents those teams already use. It began with an AI code review agent and now leads with AI Architect, a context engine for coding agents, and Governor, a gateway that routes their model calls.

AI Architect builds a knowledge graph of a company's code, Git history, Jira or Linear tickets, Confluence pages, Slack threads and observability data, keeps it current without manual reindexing, and serves it over MCP to Claude Code, Cursor, Windsurf, Copilot, Junie, Claude and ChatGPT through 13 search and lookup tools. It also acts on its own: it triages new Jira epics and stories, posts a plan as a ticket comment, and once the plan is approved implements the change on a new branch and opens a pull request linked to the ticket, and it can do the same from a Slack thread.

Twelve installable agent skills, from feasibility studies and PRDs to production triage and coding task execution, package it for other agents. The AI Code Review Agent reviews every pull request on GitHub, GitLab and Bitbucket, posts inline comments and summaries, runs static analyzers such as ESLint, Ruff and Snyk, applies one click fixes as new commits, and learns rules from negative feedback. Governor is a drop in base URL that adds AI Architect context to each coding agent request and routes it to a right sized model across four tiers on the customer's own provider keys. Orchestra, a fleet of autonomous agents, is in closed beta.

Code review costs $12 per seat a month billed annually on Team ($15 monthly), with 5,000 lines per seat a month, or $20 on Professional ($25 monthly) with a 14 day free trial; Enterprise is custom, and AI Architect and Governor are priced through sales by the size of the indexed codebase. Bito holds SOC 2 Type II, offers SSO on Enterprise, and says code and AI requests are never used to train models. The cloud service runs in the US, and Governor, AI Architect and code review can all be self hosted.

It fits engineering organizations running several coding agents that want shared codebase context, ticket driven planning and pull request review from one vendor, including those that need code kept inside their own infrastructure; named customers include PubMatic, Privado, Kredivo and Gainsight. It fits less well for teams on Azure DevOps, which only the self hosted AI Architect supports, and for buyers who need a full audit trail of every action an agent took inside Bito itself.

Vendor details

Canonical URL

https://bito.ai

Subcategory

Codebase context engine, model gateway and pull request review agent that work alongside a team's coding agents

Funding status

Private company. Bito Inc. of Menlo Park, California, founded in 2021 by Amar Goel, Anand Das and Mukesh Agarwal.

Company status

independent

Use cases & customers

Primary use cases

serving codebase, ticket and doc context to coding agents over MCPturning approved Jira plans into branches and pull requestsautomatic pull request review with fixes and learned rulesrouting coding agent requests to right sized models on the customer's keystechnical design, PRDs and feasibility studies grounded in the codebase

Target customers

engineering organizations running several coding agentsplatform and developer experience teams managing model spendteams wanting automated pull request review on GitHub, GitLab or Bitbucketcompanies that need coding agent tooling self hosted

Deployment options

SaaS (US)Self hosted on Docker or KubernetesOn premises (Enterprise)

Integrations

Code review runs on GitHub, GitLab and Bitbucket, cloud or self managed, with static analyzers such as ESLint, Ruff, Mypy, golangci-lint and Snyk. AI Architect draws on Jira, Linear, Confluence, Google Docs, Slack and observability data, supports Azure DevOps when self hosted, and serves Claude Code, Cursor, Windsurf, Copilot, Junie, JetBrains AI Assistant, Claude and ChatGPT over MCP. Governor works with Claude Code, Cursor, Cline, Continue, Aider, Codex CLI, Copilot CLI, Windsurf, Zed and any OpenAI or Anthropic compatible tool. IDE extensions cover VS Code, JetBrains, Cursor, Windsurf and Vim.

In practice

A product manager files a Jira story for a new export option. AI Architect scores it, posts an implementation plan as a comment, and after the tech lead approves it writes the change on a new branch and opens a pull request linked to the ticket.

A developer opens a pull request on GitLab. Bito's review agent flags a missing null check and an unused import, the developer applies the null check fix with one click, and marks the import comment unhelpful, which teaches the agent to stop raising it after three such signals.

A platform team points Claude Code and Cursor at Governor's base URL. Simple edits go to a small model, complex refactors to a frontier model, and the team watches the routing in shadow mode for a week before switching it on.

Sources & related URLs

Related / legacy domains

https://docs.bito.ai https://trust.bito.ai https://github.com/gitbito

Research sources

Agentic Index coverage score

11.5 / 14 capabilities · 82%

Integrations & Tool Calling	Full
Bito's agents act inside the tools teams already run. The code review agent posts inline comments, summaries and change requests on GitHub, GitLab and Bitbucket pull requests, cloud or self managed, and commits one click fixes back to the branch. AI Architect posts plans as Jira comments, creates branches and opens pull requests linked to the ticket, and works from Slack threads. It draws on Jira, Linear, Confluence, Google Docs and observability data, and reviews run static analyzers such as ESLint, Ruff, Mypy and Snyk. Sourcedocs.bito.ai/ai-architect/key-features.mdread 2026-11-08
Workflow Orchestration	Full
AI Architect runs a ticket to pull request pipeline: it scores new or updated Jira epics and stories from 1 to 10, plans those above the threshold, posts the plan for approval, then implements it on a new branch and opens a pull request linked back to the ticket. Named platform agents for issue analysis, code review and Slack split the work, and skills such as plan to coding tasks and coding task executor break a plan into steps for other agents. Orchestra, a fleet of agents that plans, builds, tests, reviews and merges, is in closed beta. Sourcedocs.bito.ai/ai-architect/key-features.mdread 2026-11-08
Knowledge Grounding & RAG	Full
AI Architect keeps a knowledge graph of the customer's code, Git history, Jira or Linear tickets, Confluence pages, Slack threads, observability data and custom instructions, tracking more than 160 data points across 15 dimensions for each repository and updating without manual reindexing. Coding agents query it over MCP with tools for searching repositories, symbols, code and fields across services. Hosted, Bito stores summaries and indexes of the code, not the code itself; self hosted, everything stays local. Reviews add symbol indexes, syntax trees and embeddings. Sourcedocs.bito.ai/ai-architect/knowledge-graph.mdread 2026-11-08
Human Oversight & Guardrails	Full
People approve before code changes. AI Architect posts its plan to the Jira ticket and implements it only after approval, and review fixes land only when someone clicks to apply them. Governor starts each new router in shadow mode, where routing decisions are logged without taking effect, routes downward only by default and holds a price ceiling. Change request comments can block a merge through the Git provider's own rules, while a request made in Slack can go straight to a branch and pull request. Sourcedocs.bito.ai/governor/auto-ai-model-routing.mdread 2026-11-08
Security, Identity & Governance	Full
Workspaces have Owner, Admin and Member roles, and Enterprise adds single sign on, with AI Architect supporting SAML 2.0 and OIDC through Descope and adding new users on first sign in. Bito holds SOC 2 Type II, with the report shared on request, and covers GDPR and CCPA. AI requests are not retained, the review diff and repository copy are deleted after each review, production logs are kept for 10 days, and Bito says code and AI requests are never used to train models. Sourcedocs.bito.ai/privacy-and-security.mdread 2026-11-08
Observability & Auditability	Partial
Governor shows 30 days of spend by model, key or provider and keeps a log of each request with CSV export. Code Review Analytics reports by submitter, repository and pull request and exports to PowerPoint or PDF, and MCP Analytics counts tool use across the workspace. The review comments and AI Architect plans themselves live in Git and Jira, and Bito keeps no audit log of its own recording each action an agent took and why. Sourcedocs.bito.ai/ai-code-reviews-in-git/code-review-analytics.mdread 2026-11-08
Memory & State Persistence	Partial
The code review agent learns from feedback. A negative reaction to a comment creates a rule, held off at first and switched on automatically after three negative signals, applied at the repository level for that programming language, and teams can switch any learned rule on or off. AI Architect folds corrections given in Slack or Jira back into its knowledge graph so they apply for everyone on the team. Neither store has a stated retention period or a way to clear what was learned. Sourcedocs.bito.ai/ai-code-reviews-in-git/implementing-custom-code-review-rules.mdread 2026-11-08
Deployment & Data Residency	Full
The hosted service runs in the United States. Governor can be self hosted on Docker or on Kubernetes with Helm, AI Architect runs self hosted in standalone or enterprise mode, code review self hosting is a $5 per seat add on for Professional plans, and on premises deployment comes with Enterprise. Governor forwards requests to the customer's own model provider accounts, including a passthrough mode, and self hosted AI Architect works with Anthropic, OpenAI, Gemini, Azure AI, Bedrock, Vertex or Portkey. Sourcedocs.bito.ai/ai-architect/installation/install-ai-architect-self-hosted/custom-llm-configuration.mdread 2026-11-08
Prebuilt Agents / Templates / Packs	Full
Twelve agent skills install into Claude Code, Cursor, Windsurf, Copilot and Junie: codebase explorer, build plan, feasibility, spike, feature plan, PRD, TRD, production triage, code review, commit review, plan to coding tasks and coding task executor. Each does a job of its own on top of the knowledge graph. Bito also ships named agents for issue analysis, pull request review and Slack, ready to switch on for a workspace. Sourcedocs.bito.ai/ai-architect/agent-skills.mdread 2026-11-08
Triggers & Channel Coverage	Full
Work starts from events across the delivery chain. Every new pull request gets a review automatically, with incremental reviews on each push, and comment commands such as /review with a security or performance scope, /pause, /resume and /abort steer it. CI scripts run reviews from Jenkins, Argo CD, GitLab CI or GitHub Actions. New and updated Jira epics and stories are triaged on their own, and an @bito mention, a bito label or a Slack message starts AI Architect. Sourcedocs.bito.ai/ai-code-reviews-in-git/available-commands.mdread 2026-11-08
Model Flexibility & Routing	Full
Customers choose the models. Governor sorts each coding agent request into simple, medium, complex or reasoning tiers and sends it to the model the customer assigned to that tier, under a quality, balanced or cost mode, across Anthropic, OpenAI and Groq accounts on the customer's own keys, with models such as Opus 5.5, Sonnet 5.5, Haiku 5.5 and GLM 5.3 Flash. Self hosted AI Architect takes a custom model configuration across seven providers and gateways. Sourcedocs.bito.ai/governor/auto-ai-model-routing.mdread 2026-11-08
APIs / SDKs / MCP Extensibility	Full
Governor exposes OpenAI and Anthropic compatible endpoints for messages, chat completions and responses at gateway.bito.ai, with an admin command line tool for managing routers and keys. AI Architect runs an MCP server for each workspace that authenticates with a bearer token or single sign on and offers 13 tools for listing and searching repositories, symbols, code and fields. The bitoreview command line tool and CI scripts drive reviews, though no general REST API manages Bito itself. Sourcedocs.bito.ai/ai-architect/available-mcp-tools.mdread 2026-11-08
Testing, Debugging & Optimization	Partial
Governor's shadow mode lets a team watch what a routing policy would do before it takes effect, and code review has a test mode for trying configurations. Bito publishes benchmark results of its own, including a public software engineering benchmark run with and without AI Architect and a cost study of 29 models on 60 tasks. There is no harness for scoring a customer's agents on saved test cases or for gating a release on the results. Sourcedocs.bito.ai/governor/auto-ai-model-routing.mdread 2026-11-08
Browser / Computer-use	Not documented
Bito works through repositories, tickets, chat, model APIs and MCP. Its agents read and write code, comments and tickets, and AI Architect can research the web while drafting a design, but nothing drives a browser, a desktop or a web interface, and no hosted browser is offered. Pull request reviews, plans and fixes all travel through the Git, Jira and Slack APIs. Sourcebito.ai/product/technical-design-and-scopingread 2026-11-08

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

AI code review on Team is $12 per seat a month billed annually, or $15 monthly. Professional is $20 per seat billed annually, or $25 monthly. Enterprise is custom, and AI Architect and Governor are quoted by sales.

Per seat per month for code review, monthly or annual, with a reviewed line allowance on Team; AI Architect and Governor priced on the size of the indexed codebase.

Free tierTrial available

Included quota

Team includes 5,000 reviewed lines per seat a month for up to 25 seats. Professional has unlimited seats.

What is public

Per seat prices for Team and Professional code review, monthly and annual, the Team line allowance and overage rate, the self hosting add on, and the free trial terms.

Billing mechanics

Code review seats bill monthly or annually by card. AI Architect, Governor and Enterprise are contracted through sales.

Cost watchouts

Model spend routed through Governor is billed by the customer's own providers. Self hosting code review adds $5 per seat on Professional. Team stops at 25 seats.

Variable cost rationale

Code review is priced per seat, with Team capped at 5,000 reviewed lines per seat a month and $5 for each extra 1,000 lines, so heavy review volume raises the bill. Governor and AI Architect scale with the size of the indexed codebase, and Governor routes calls onto the customer's own model accounts, where model spend sits.

Additional watchouts

AI Architect and Governor, the products Bito now leads with, carry no public rates. Team is limited to 25 seats and a monthly line allowance.

Overage / add-ons

On Team, each 1,000 reviewed lines past the seat allowance costs $5.

Sales call required

Mixed (some tiers require a call)

Free / trial

A free tier gives pull request summaries. Professional comes with a 14 day free trial and no card, and AI Architect and Governor offer free trials.

Lowest paid plan

Team at $12 per seat a month billed annually, or $15 monthly, for up to 25 seats

Commercial notes

Bito is sold by Bito Inc. of Menlo Park, California, founded in 2021. Named customers include PubMatic, Privado, Kredivo, Apica and Gainsight.

Key ambiguities

Governor is described as usage based and scaling with the size of the indexed codebase, without a stated unit or rate.

Cancellation / refund

Monthly or annual billing for code review seats; Enterprise terms by contract.

Support SLA / resale

Enterprise adds SSO and on premises deployment; support levels for other plans are not stated.

Missing data

Rates for AI Architect and Governor, and Enterprise pricing.

Agentic Index verified 2026-11-08

Official pricing page Docs

Data confidence: high