How much does Tempo Labs cost?

Free includes 250 issues, 100 docs, 500 local tests a month and seven days of history. Pro is $50 a month for one user, Team is $200 a month for five users, extra users cost $40 a month each, and Enterprise is custom. The monthly price counts toward model usage billed at break even API rates, with top ups, and extra tests cost $4.50 per 1,000 on Pro. The separate Agent+ service starts at $500 a month.

Does Tempo Labs give you tracing and observability into agent runs?

Full under the Agentic Index verification standard. Tempo Labs publishes public evidence of tracing, observability and audit logging into agent runs, so a team can see what an agent did and why. Evidence basis: Every agent run is kept as a chat under that agent that people can watch live, scroll back through or join mid run, and runs link to the artifacts they produce. Analytics show agent chat usage, a plan and usage page tracks spend, a daily di. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does Tempo Labs require human approval before an agent acts?

Full under the Agentic Index verification standard. Tempo Labs publishes public evidence of human approval steps and guardrails, so an agent action can be held for review before it executes. Evidence basis: Agents are created disabled and switched on by a person, and template agents are never created or enabled on their own. Agent work lands as a pull request with its issue card moved to review, agents can stop to ask approval questions that s. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Can you bring your own model or route between providers with Tempo Labs?

Full under the Agentic Index verification standard. Tempo Labs publishes public evidence of bring your own model and bring your own key (BYOK), and of routing tasks between models or providers rather than being locked to one. Evidence basis: Each agent has its own model and effort setting, and Tempo Inference bills models to the plan, from Claude Opus 5.5, Sonnet and Haiku to GPT 6 variants, Grok 4.7, Kimi K3 and DeepSeek V4 Pro, with a free small model to start. Users can add. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

Tempo Labs

Q: Who is Tempo Labs for?

startup product teams that ship web and mobile apps, product managers and designers who work in the same tool as engineers, teams with a React, Vue or React Native component library, engineers who already use Claude Code or Codex

Also known as: Tempo Labs Inc., Tempo IDE, Tempo Agent+, tempo.new

Visit site

Coding agentindependentVerified 2026-11-08

Entry priceFree with 250 issues and 100 docs, or Pro at $50 a month for one user, with the monthly price counting toward model usage.Full pricing detail

Tempo Labs makes Tempo, a desktop IDE for macOS, Windows and Linux where a whole product team writes docs, designs on canvases built from the codebase's real components and runs agents that triage feedback, write specs, fix bugs and build features as pull requests. Tempo's own agents start on assignment, issue events, schedules, Slack messages or GitHub events, and it also hosts Claude Code and Codex on the customer's subscription. Pro costs $50 a month, applied to inference at cost.

Tempo Labs is a Toronto company, Tempo Inc. in its terms, founded in 2023 by Kevin Michael and Peter Gokhshteyn and backed by Y Combinator from its Summer 2023 batch. It began as a visual editor for React apps on the web, wound that editor down in June 2026 and now sells Tempo as a desktop app for macOS, Windows and Linux, described as the IDE for the whole team. Separately, its Agent+ service has Tempo's agents design and ship features with human engineers and designers reviewing the work, priced by the month.

Inside the app a team keeps docs, an issues board, chat and canvases, where each frame is a live render of a real route or component from the repository, stored as code and shipped as a pull request. Tempo's own agents are standing prompts wired to triggers: an issue assigned to the agent, issue events, a schedule, a Slack message or mention, a GitHub event or a published release.

A seeded fleet covers triage of feedback into prioritized issues, specs and clickable designs, bug fixing and feature building, and each agent works in its own git worktree, reads the linked docs, canvases and issues, changes code and opens a pull request that moves the card to review. Agents can hand work to a teammate's agent, post to Slack, read and write Linear issues, and run on a model and effort level set per agent. Chat sessions can also run Claude Code, Codex or opencode on the customer's own subscription.

Agents start disabled and are switched on by hand, and a Visual Review check on GitHub compares tracked canvases pixel by pixel on every pull request and stays red until a person approves. Every agent run is kept as a chat that can be watched live or replayed, with history kept for 7, 90 or 365 days by plan. Agents run on the customer's machine while the app is open, while docs, issues, canvases and chat history sync to Tempo's cloud. Tempo names no security certification or single sign on, and its terms let it use customer content to improve or train its product. An MCP server and CLI let other agents read and write Tempo issues, docs and agents.

Free includes 250 issues, 100 docs and seven days of history, Pro costs $50 a month for one user and Team $200 a month for five, with $40 for each extra user, and the monthly price counts toward model usage billed at cost. Enterprise adds bring your own infrastructure and unlimited history with audit. It fits product teams that want specs, designs and code changes in one tool with agents working from their real components, and startups that want agents running on their own machines. It fits less well for buyers who need certifications, single sign on, agents that run while laptops are closed, or a public API.

Vendor details

Canonical URL

https://www.tempo.new

Subcategory

Desktop IDE for product teams whose own triggered agents triage feedback, write specs and designs, fix bugs and build features as pull requests, beside hosted Claude Code and Codex

Funding status

Backed by Y Combinator from its Summer 2023 batch; no funding amounts are stated on Tempo's own pages.

Company status

independent

Use cases & customers

Primary use cases

turning Slack feedback into prioritized issueswriting specs and clickable designs from real componentsfixing bugs and building features as pull requestsscheduled and event triggered agent runsvisual regression review of canvases on pull requests

Target customers

startup product teams that ship web and mobile appsproduct managers and designers who work in the same tool as engineersteams with a React, Vue or React Native component libraryengineers who already use Claude Code or Codex

Deployment options

Desktop app for macOS, Windows and Linux, with agents running on the user's machineRemote Control to run workspaces on another of the user's machinesTempo cloud for organization docs, issues, canvases and chat historyEnterprise bring your own infrastructure

Integrations

A GitHub App for repositories, pull requests, commit statuses and webhook triggers, with GitLab remotes and Bitbucket Cloud also supported; a Slack bot that agents post through, with direct messages and thread replies; Linear issue mirroring with read and write; Figma paste into editable canvases; Claude Code, Codex and opencode on the customer's own logins; and an MCP server and CLI that expose Tempo issues, docs, comments, agents, scripts, Slack and Linear to other agents.

In practice

A product manager at a startup lets the Triage agent watch the team's feedback channel in Slack. Each bug report becomes a prioritized issue, the Spec and Design agent drafts a PRD with a clickable canvas built from the app's components, and once an engineer assigns the issue to the Feature Builder agent it builds the change on its own branch and opens a pull request.

A designer adjusts spacing on a canvas that renders the real checkout route, the edit writes back to the source, and the Visual Review check on the resulting pull request shows the pixel difference and waits for her approval before merging.

An engineer schedules an agent to run each morning, read new Linear issues tagged as bugs, attempt fixes in separate worktrees and post a summary to Slack, while she works in a parallel Claude Code session on her own subscription.

Sources & related URLs

Related / legacy domains

https://docs.tempo.new https://www.tempo.new/pricing https://www.tempo.new/ai-agent-plus https://docs.tempo.new/changelog

Research sources

Agentic Index coverage score

11.5 / 14 capabilities · 82%

Integrations & Tool Calling	Full
Through its GitHub App, Tempo connects repositories, opens pull requests, posts commit statuses and accepts repository webhooks, and projects can also use GitLab remotes and Bitbucket Cloud. Agents post to Slack as the Tempo bot across several workspaces with 13 Slack tools including direct messages, and replies in a thread return to the same session. Linear issues mirror into the organization and can be read and written, and Figma designs paste in as editable canvases. No Jira connector exists. Sourcedocs.tempo.new/changelogread 2026-11-08
Workflow Orchestration	Full
Each agent works in its own git worktree, so several run against one repository without colliding, and a default Software Factory pipeline carries work through stages from triage to spec, build and review. Agents can route work to a teammate's agent, people delegate to agents by mentioning them in chat, and subagent output appears in agent cards. Remote Control moves a workspace onto another of the user's machines, and multiple repositories work together through git submodules. Sourcedocs.tempo.new/product/agentsread 2026-11-08
Knowledge Grounding & RAG	Full
Every turn, an agent sees its worktree's file tree and git history together with the docs, canvases and issues linked to the work, and chat can read the live workspace. Docs persist in the organization's database as a searchable store beside the codebase, components and token libraries come from the repository into the canvas, assets and stylesheets are indexed, and design context can be pulled from external MCP servers. Tempo describes no semantic or vector index of the code itself. Sourcedocs.tempo.new/product/docsread 2026-11-08
Human Oversight & Guardrails	Full
Agents are created disabled and switched on by a person, and template agents are never created or enabled on their own. Agent work lands as a pull request with its issue card moved to review, agents can stop to ask approval questions that show the configuration they refer to and wait for the answer, organization admins approve access requests, and the MCP server can be limited to read only use for any client that connects to it. Sourcedocs.tempo.new/product/agentsread 2026-11-08
Security, Identity & Governance	Partial
Organizations have two roles, Admin and Member, members can invite but not remove others, and each agent can be private or public with view only or edit rights, while agents and MCP tokens act as the user who owns them. Tempo names no SOC 2, ISO or other certification, no single sign on and no encryption detail, and its terms allow customer content to be shared with vendors to improve or train the product or those vendors' services. Sourcedocs.tempo.new/legal/termsread 2026-11-08
Observability & Auditability	Full
Every agent run is kept as a chat under that agent that people can watch live, scroll back through or join mid run, and runs link to the artifacts they produce. Analytics show agent chat usage, a plan and usage page tracks spend, a daily digest summarizes pull request reviews, and history is kept for 7, 90 or 365 days by plan, with unlimited history and audit on Enterprise and a public status page for the service. Sourcedocs.tempo.new/product/agentsread 2026-11-08
Memory & State Persistence	Partial
Chat history persists per workspace, Slack threads continue the same session, and shared setup and run scripts live in a project config file in the repository. A custom system prompt applies to every conversation and personal skills carry across chat providers, but these act as standing instructions, and no memory that agents build from past runs or corrections is described anywhere in the product. Sourcedocs.tempo.new/changelogread 2026-11-08
Deployment & Data Residency	Partial
Agents run on the customer's own machine and only while the desktop app is open, and Remote Control can move a workspace to another machine the user owns. Organization docs, issues, canvases and chat history live in Tempo's cloud database, with processing in Canada, the United States and elsewhere and no region choice, and bring your own infrastructure appears only as one line on the Enterprise plan. Sourcedocs.tempo.new/product/agentsread 2026-11-08
Prebuilt Agents / Templates / Packs	Full
New organizations start with a seeded fleet of template agents, including Triage, Spec and Design, Bug Fixer, Feature Builder, pull request review and release driving, each carrying a diagram of the pipeline it fits. Built in slash commands such as a PRD writer and a test driven development audit, a design system command suite, a visual review workflow template and a web clipper give teams ready starting points. Sourcedocs.tempo.new/product/agentsread 2026-11-08
Triggers & Channel Coverage	Full
An agent starts when an issue is assigned to it, when issue events fire, on a cron schedule in the user's time zone, on a Slack message or mention, on a GitHub repository event through a webhook, or when a release is published, and missed scheduled runs catch up later. People can also wake an agent by mentioning it in chat or answering one of its approval questions. Every trigger depends on the desktop app running, since nothing runs while Tempo is closed. Sourcedocs.tempo.new/product/agentsread 2026-11-08
Model Flexibility & Routing	Full
Each agent has its own model and effort setting, and Tempo Inference bills models to the plan, from Claude Opus 5.5, Sonnet and Haiku to GPT 6 variants, Grok 4.7, Kimi K3 and DeepSeek V4 Pro, with a free small model to start. Users can add their own provider API keys per device, and chat sessions can instead run Claude Code, Codex or opencode on the customer's own logins. Sourcedocs.tempo.new/changelogread 2026-11-08
APIs / SDKs / MCP Extensibility	Partial
A hosted MCP server over streamable HTTP takes OAuth sign in, and an npm package runs the same tools as a local MCP server or a command line with personal access tokens, with toolsets for issues, docs, comments, agents, scripts, Slack, Linear and canvases and a read only switch. Tempo names the toolsets but lists only a few tools by name, and there is no public REST API or SDK. Sourcedocs.tempo.new/guides/install-mcpread 2026-11-08
Testing, Debugging & Optimization	Full
Visual Review compares tracked canvases pixel by pixel on every pull request through GitHub Actions, after setup confirms each canvas renders identically three times, and posts a required check on the result. A built in command audits test driven development, user defined run scripts can run a project's tests, and a checks panel shows results beside each change, while the Bug Fixer agent is described as shipping its fix as a tested pull request. Sourcedocs.tempo.new/guides/visual-reviewread 2026-11-08
Browser / Computer-use	Partial
Canvases render real routes as live frames, MCP screenshots render headlessly in Chromium, and the app has embedded browser tabs with element context for people to point at. The macOS install asks for Full Disk Access and Accessibility permissions, and a recent release patched browser automation security issues, but no Tempo agent is described driving a browser or desktop to reproduce a bug or click through a feature. Sourcedocs.tempo.new/changelogread 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

Free with 250 issues and 100 docs, or Pro at $50 a month for one user, with the monthly price counting toward model usage.

Monthly plans with included users, where the subscription price is credit toward model usage billed at break even API rates, plus extra users and top ups.

Free tier

Included quota

Pro includes one user, unlimited issues and docs, 5,000 tests a month and 90 days of history; Team includes five users, 35,000 tests a month and 365 days of history.

What is public

Free, Pro and Team prices, per user add ons and test overage rates, with Enterprise and the Agent+ service by contact.

Billing mechanics

Monthly plans whose price is credit toward model usage at cost, with included users, test allowances, top ups and extra users.

Cost watchouts

Agent runs on large models draw down the plan's dollar allowance quickly, and the docs advise base plans to use smaller models at low effort.

Variable cost rationale

The plan price doubles as a model usage allowance at cost, so heavy agent runs on larger models use it up and need top ups, though customers can shift work to their own Claude or Codex subscriptions or their own API keys.

Additional watchouts

No annual price or trial, and plan units on the pricing page differ from the usage based billing in the docs.

Overage / add-ons

Model usage beyond the plan's dollar amount needs top ups at break even API rates, extra tests cost $4.50 per 1,000 on Pro, and extra users cost $40 a month each, up to 49 more.

Sales call required

Mixed (some tiers require a call)

Free / trial

The free plan includes 250 issues, 100 docs, 500 local tests a month and seven days of history, with model usage metered by spend on a starter model.

Lowest paid plan

Pro at $50 a month for one user

Commercial notes

Tempo Labs sells the desktop app directly and separately sells Agent+, a monthly design and development service run by its agents and staff.

Key ambiguities

The pricing page counts issues, docs and tests, which no docs page defines, while the billing guide describes plans as dollar credit for model usage.

Cancellation / refund

Monthly subscriptions; refunds only for verified technical issues requested within seven days, while the separate Agent+ service offers first month money back.

Support SLA / resale

Team adds more users, tests and history, and Enterprise adds bring your own infrastructure, custom test volume and unlimited history with audit through a sales call.

Missing data

Enterprise pricing, a definition of tests and the size of the free plan's usage allowance.

Agentic Index verified 2026-11-08

Official pricing page Docs

Data confidence: medium