Back to vendors
T

Tabby

Also known as: Pochi

Visit site
Entry priceCommunity is free and open source for up to 5 users with local deployment; Team is $19 a month per seat for up to 50 users with flexible deployment; Enterprise is custom with unlimited users, SSO and group managementFull pricing detail

Open source, self hosted coding assistant serving fill in the middle completion, an Answer Engine for codebase questions, inline chat and a code browser on the customer's own hardware, with open models the customer chooses and no external database or cloud service. TabbyML also makes Pochi, an open source coding agent for VS Code that runs parallel agents, drives a browser and calls MCP tools.

Tabby is an open source, self hosted AI coding assistant built by TabbyML. Code and inference stay on the customer's own infrastructure. It is written in Rust, distributed under Apache 2.0, and runs as a single self contained server installed through Docker, a standalone binary, Homebrew or Kubernetes. It needs no external database or cloud service. The server loads an open coding model onto a GPU and serves suggestions over a REST interface described in OpenAPI, and it runs on consumer grade NVIDIA, Apple and AMD hardware. The project has a long release history and more than a hundred contributors.

Its core is code completion tuned for the fill in the middle pattern that autocomplete relies on. Around that, Tabby adds an Answer Engine that lets a developer ask questions about the codebase and get grounded explanations in the editor, an inline chat for real time back and forth, and a code browser. The server keeps a persistent index of the team's repositories, parsed with Tree-sitter and embedded. Context Providers pull in documentation, configuration files, Git repositories and external interfaces, and completions, the Answer Engine and chat all retrieve from that index. Answer Engine threads can be turned into persistent, shareable Pages for a team.

Tabby is model agnostic. Because the server is self hosted, the customer supplies and runs the weights. The model is chosen at launch, a separate chat model can be paired with the completion model, and embeddings are configurable independently. Supported families include StarCoder and StarCoder2, CodeLlama, Qwen and Qwen2.5-Coder, and Codestral through a Mistral integration, with Mistral embeddings as well. Llamafile deployment is supported, and teams can fine tune on their own code. Tabby integrates as a plugin across VS Code, the JetBrains family, Vim, Neovim and Emacs, and works with cloud editors.

The Tabby server serves completions, answers and chat. It does not plan, execute or act on external systems. That work belongs to Pochi, TabbyML's open source coding agent, which installs from the VS Code Marketplace and OpenVSX under a Research Preview label and also runs as a CLI and a GitHub Action. Pochi runs several agents at once in separate Git worktrees and hands parts of a task to subagents. It calls tools on MCP servers, drives a managed browser or the developer's own Chrome, and keeps a project memory file across tasks.

Each type of action can be auto approved or held for the developer, who then sees the exact diff before anything is written. Pochi runs on the developer's own key for any OpenAI compatible provider, on local models or on managed models through a Pochi account. That cloud account, which adds task storage, sharing and team features and is needed for the GitHub Action, is still offered through a waitlist.

For administration, Tabby ships an admin dashboard for user accounts, API tokens per developer, model management and usage analytics. Team deployments can use LDAP directory authentication and generic OAuth. SSO and group management sit on the Enterprise plan. TabbyML names no security certification.

The Community plan is free and open source for up to 5 users on local deployment. Team is $19 a month per seat for up to 50 users, and Enterprise is custom with unlimited users. On every plan the team also pays for the hardware or cloud compute that serves the models and needs the capacity to run and maintain model serving. Pochi bills the token cost of the models it runs, with $20 of free credit a month.

Vendor details

Canonical URL

https://tabbyml.com

Category

Coding agent

Subcategory

Open source self-hosted coding assistant (completion, answer engine, chat)

Funding status

TabbyML is the company behind the open-source Tabby project, distributed under Apache 2.0. Specific venture funding is not disclosed on first-party sources. The project reports roughly 33,800 GitHub stars and 1,780 forks with over a hundred contributors, and the latest stable release documented is v0.32.0 on 25 January 2026, adding Mistral Embedding API support, generic OAuth and multi-branch repository indexing. Note that the tabby repository's last recorded activity is 30 June 2026, while the company's newer TypeScript project Pochi was updated 24 August 2026, suggesting engineering attention has shifted.

Company status

independent

Use cases & customers

Primary use cases

private self hosted code completioncodebase question answering in the IDEAI assistance for regulated and air gapped teamsmodel agnostic coding assistant on owned hardware

Target customers

privacy first and regulated teams (finance, healthcare, defense)organizations that run their own infrastructureteams wanting an open source alternative to cloud coding assistants

Deployment options

Open-source self-hosted (Docker, standalone binary, Homebrew)On-premisesPrivate cloud / KubernetesConsumer-grade GPUs (NVIDIA CUDA, Apple Metal, AMD ROCm)Fully offline, no external DBMS or cloud

Integrations

Tabby runs as a self hosted server exposing a REST API for completion, chat and answers. It connects to IDEs through official plugins for VS Code, the JetBrains family, Vim, Neovim and Emacs, plus cloud editors. Context Providers ingest documentation, configuration files, Git repositories and external interfaces to ground the assistant, and custom documentation can be added over the API. An admin dashboard manages users, models and usage. Tabby serves completion, answers and chat and does not act on external tools, and TabbyML names no Tabby agent capability.

In practice

You cannot send source code to a cloud service. Tabby runs entirely on your own hardware or private cloud, serving completion, chat, and answers with no external calls, so code never leaves your infrastructure.

You want to pick your own model. Tabby is model agnostic, running open coders like StarCoder, CodeLlama, or Qwen, and you can combine a completion model with a chat model or fine tune on your codebase.

Your small team wants AI coding help without a subscription. Tabby's Community plan is free and open source for up to 5 users on local deployment, so the main cost is the hardware or cloud compute running the server.

Agentic Index coverage score

11.0 / 14 capabilities · 79%

Integrations & Tool Calling Full

Pochi connects to MCP servers, local over stdio or remote over HTTP and SSE, with credentials passed as request headers. Servers are set globally or per project, and each tool can be switched on or off. A Use MCP permission governs when the agent calls databases, APIs and other services through them. Through its GitHub Action, Pochi also comments on and commits to the repository. The self hosted Tabby server takes in context through Context Providers and REST APIs and returns suggestions.

Sourcedocs.getpochi.com/mcp and /permissionsread 2026-10-25

Workflow Orchestration Full

Pochi, TabbyML's coding agent, runs several agents at once, each in its own Git worktree, and each task opens in its own tab with a diff against main. Since September 2026 a task can also hand independent parts to multiple subagents that work while the main conversation continues. Custom agents can launch other agents as new tasks. Pochi carries a Research Preview label but installs publicly from the VS Code Marketplace and OpenVSX. The self hosted Tabby server itself serves single turn completion, answers and chat.

Sourcedocs.getpochi.com/parallel-agents and /developer-updatesread 2026-10-25

Knowledge Grounding & RAG Full

The server keeps a persistent index of the customer's repositories, parsed with Tree-sitter and embedded. Background jobs reindex it, and v0.32.0 extended it to multiple branches. Context Providers, which are data connectors, add documentation, configuration files, Git repositories and external interfaces, and custom documentation can be added over REST APIs. Completions, the Answer Engine and chat all retrieve from that index.

Sourcegithub.com/TabbyML/tabby release feed and tabbyml.comread 2026-09-29

Human Oversight & Guardrails Full

Pochi sorts its actions into Read, Write, Execute, Use MCP and Retry permissions, and each can be auto approved or left to the developer. With approval on, the developer sees the exact change before approving it, and nothing is written to disk until then. All five are auto approved by default. In the self hosted Tabby server, the developer accepts or rejects every completion, chat answer and inline edit.

Sourcedocs.getpochi.com/permissions and /developer-updatesread 2026-10-25

Security, Identity & Governance Partial

Team deployments support LDAP directory authentication, shipped in v0.24.0, and generic OAuth, added in v0.32.0. An admin dashboard manages user accounts, issues API tokens per developer and handles model configuration. Because Tabby is self hosted, code and inference never leave the customer's infrastructure and no external service holds data. Tabby names no security attestation or certification, SAML, SCIM, fine grained role based access control or audit logging.

Sourcegithub.com/TabbyML/tabby and tabby.tabbyml.com/docsread 2026-08-30

Observability & Auditability Partial

Every Pochi run is saved as a task, stored locally or in the cloud for signed in users, and can be shared by link. Completed steps show how long they took, and browser sessions can be recorded. No admin view or audit log across a team's tasks is described. The self hosted Tabby server has an admin dashboard with user accounts, per developer API tokens and usage analytics.

Sourcedocs.getpochi.com/vscode and /developer-updatesread 2026-10-25

Memory & State Persistence Partial

Pochi keeps a project memory file, MEMORY.md, that it writes and carries forward across tasks, on by default since May 2026. A separate task memory file survives context compaction within a task. No lifetime or retention is stated for the project memory. In the self hosted Tabby server, Answer Engine threads persist as shareable Pages and repository indexes persist between sessions.

Sourcedocs.getpochi.com/developer-updatesread 2026-10-25

Deployment & Data Residency Full

The code is written in Rust, and Tabby is distributed under Apache 2.0 as a self contained server that needs no external database management system or cloud service. It installs by Docker, standalone binary, Homebrew tap or Kubernetes, and runs on consumer grade hardware across NVIDIA CUDA, Apple Metal and CPU. Teams without local hardware can deploy it on GPU cloud hosts. Tabby runs on the customer's own terms, in the cloud or on premises, so code and inference stay inside the customer boundary.

Sourcegithub.com/TabbyML/tabby and tabbyml.comread 2026-08-30

Prebuilt Agents, Templates & Packs Partial

Pochi loads skills from project and global folders, including the shared .agents folder that other coding agents also read. Community skills install from the third party skills.sh registry. Developers define custom agents with their own prompts and toolsets. TabbyML publishes no catalog of its own agents, templates or skills, and the self hosted Tabby server ships one assistant.

Sourcedocs.getpochi.com/skills and tabbyml.com/agentread 2026-10-25

Triggers & Channel Coverage Partial

Pochi's GitHub Action runs when a comment starts with /pochi, and workflows can post that comment on new pull requests, pushes, new issues and a cron schedule. Pochi then reviews, fixes and commits, and its verdict can gate tests and deploys. The Action needs a Pochi API key from app.getpochi.com, the cloud account TabbyML still offers through a Research Preview waitlist. Inside a running task, Pochi can watch log messages or CI checks and resume when a new event arrives. Developers reach the self hosted Tabby server through plugins for VS Code, JetBrains, Vim, Neovim and Emacs.

Sourcedocs.getpochi.com/tutorials/pochi-github-actions and tabbyml.com/agentread 2026-10-25

Model Flexibility & Routing Full

The customer selects the model at launch through a serve flag, and a separate chat model can be paired with the completion model in the same command. Embeddings are configurable independently. Supported families include StarCoder and StarCoder2, CodeLlama, Qwen and Qwen2.5-Coder variants, and Codestral through a Mistral integration.

Mistral Embedding API support arrived in v0.32.0, alongside Llamafile deployment integration, and teams can fine tune on their own code. Because the server is self hosted, the customer supplies and runs the weights directly. Tabby names no routing gateway that sends each task to a different provider.

Sourcegithub.com/TabbyML/tabby and tabby.tabbyml.com/docsread 2026-08-30

APIs, SDKs & MCP Extensibility Full

The self hosted server publishes an OpenAPI reference for its completion, chat, event and health endpoints, which the official plugins and cloud IDEs consume. Since v0.29, custom documentation and context sources can be added over REST APIs. The reference is thin and labeled for an older release, and no SDK or MCP server is named. Pochi adds API keys for its GitHub Action and a CLI that runs a task from a prompt. The server and plugins are Apache 2.0 open source.

Sourcetabby.tabbyml.com/api and github.com/TabbyML/tabbyread 2026-10-25

Testing, Debugging & Optimization Partial

Before a Pochi task with open todos can finish, a read only audit subagent inspects the workspace, reading files, test output and command results. Those checks lean on the project's own tests rather than a harness of Pochi's. A GitHub Action workflow can also gate tests and deploys on Pochi's approve or reject verdict, and the Action needs the waitlisted Pochi cloud account. The self hosted Tabby server ships no test, review or evaluation tools.

Sourcedocs.getpochi.com/developer-updatesread 2026-10-25

Browser & Computer Use Full

Pochi's browser agent navigates websites, extracts content and interacts with pages, in a managed headless browser or in the developer's own running Chrome. Since May 2026 it connects to local Chrome over the Chrome DevTools Protocol and works inside that session, saved logins included. Sessions can be recorded, and the CLI drives the browser too. The self hosted Tabby server itself has no browser or code execution.

Sourcedocs.getpochi.com/developer-updatesread 2026-10-25

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

Community is free and open source for up to 5 users with local deployment; Team is $19 a month per seat for up to 50 users with flexible deployment; Enterprise is custom with unlimited users, SSO and group management

Per seat for Team, custom for Enterprise; the Community plan is free for up to 5 users, and the team still pays for the hardware that serves the models

Free tier

Included quota

The full open source product with no usage limits: code completion, the Answer Engine, inline chat, a code browser, Context Providers, an admin dashboard, and directory based authentication, all self hosted. Teams supply and run their own models and hardware. Community is free for up to 5 users, Team is $19 a month per seat for up to 50 users, and there are no per call charges.

What is public

tabbyml.com/pricing lists Community free and open source for up to 5 users with local deployment, Team at $19 a month per seat for up to 50 users with flexible deployment, and Enterprise custom with unlimited users, customized deployment, enhanced security and group management and SSO. The same page carries Tabby Cloud usage pricing for Pochi, which bills the token cost of the LLMs it runs with $20 in free monthly credits.

Billing mechanics

Free for up to 5 users; beyond that, Team is billed per seat monthly or yearly, and Enterprise is contracted. Model serving hardware is the customer's cost on every plan.

Cost watchouts

The real cost is infrastructure and operations: a graphics processor to serve models, and the engineering time to deploy, secure, and maintain the server. Larger models need more capable hardware. There is no vendor bill, but self hosting shifts cost and responsibility to the team.

Variable cost rationale

Tabby serves self hosted open models, so there is no per token vendor billing. Cost is the infrastructure that runs the server, which is largely fixed for a given hardware footprint and scales in steps as a team adds capacity, not continuously with each request. Exposure is therefore low and predictable, dominated by hardware rather than usage metered fees.

Additional watchouts

Self-hosting means the team pays to operate it on every plan: budget for graphics hardware or cloud compute and for the engineering time to deploy, secure, and maintain the server and models.

Overage / add-ons

Not applicable to the seat plans; Pochi, a separate product, bills automatically once usage passes its free credit.

Sales call required

Mixed (some tiers require a call)

Free / trial

Community plan: free and open source for up to 5 users with local deployment, covering completion, the Answer Engine, inline chat and Context Providers

Lowest paid plan

Team, $19 a month per seat, up to 50 users

Commercial notes

The self-hosted server is still Apache 2.0 open source, but team use above five users is now a paid seat plan.

Key ambiguities

Beyond the seat price, cost planning is mostly about infrastructure: which model, how much graphics memory, and how many concurrent developers a node can serve. Enterprise pricing is custom.

Cancellation / refund

Community needs no subscription or contract. Team is a per seat subscription bought online, billed monthly or yearly, and Enterprise is contracted. TabbyML publishes no cancellation or refund terms.

Missing data

Enterprise pricing is custom. The page also shows three annual plans ($180, $280 and $480 a year) with generic feature lists that do not match the rest of the page.

Agentic Index verified 2026-10-25

Alternatives to Tabby

The closest documented capability profiles to Tabby among coding agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

  • CodeGPT10.0 / 14A lighter documented profile than Tabby
  • Cosine10.0 / 14A lighter documented profile than Tabby
  • Playcode10.5 / 14Fuller documented coverage on Testing, Debugging & Optimization
  • Refact.ai12.5 / 14Fuller documented coverage on Memory & State Persistence and Prebuilt Agents, Templates & PacksTabby vs Refact.ai →
  • Builder.io10.0 / 14Fuller documented coverage on Security, Identity & Governance
  • Cline12.0 / 14Fuller documented coverage on Observability & Auditability and Prebuilt Agents, Templates & Packs

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.