CodeTogether

Overview

CodeTogether is a platform for measuring AI-assisted development and collaborating across editors. It is best suited to organizations that want to connect AI use with delivered work, alongside real-time coding and review. Its value measurement is distinctive, but teams relying on a source-control connector beyond GitHub should look elsewhere.

Key features

AI work and value measurement

CodeTogether captures AI spend and human time in IDEs, terminals, and agent sessions, then assigns delivered pull requests a price and value and calculates value per dollar. That gives teams a way to assess AI use against output, rather than treating activity or spend as the result. A single IDE plugin can track agentic command-line tools such as Claude Code, Codex, and Cursor running inside that IDE.

Collaboration across IDEs

Teams can code, debug, and review together in real time across IDEs, with participants joining from an IDE or browser. Support includes VS Code, IntelliJ, Eclipse, and Visual Studio. This makes CodeTogether useful for mixed-editor teams; it is less suited to sessions that depend on built-in voice or screen sharing, which it does not provide.

Deployment, privacy, and access

Organizations can run CodeTogether as SaaS or on-premises inside a private network. The company says it does not store code or train on customer data. On-premises deployments keep clients, analysis, and the database inside the private network. On Team plans, no code leaves the system; on Company plans, code is scanned only when value analysis is enabled and is never stored. On-premises AI analysis can connect to Ollama, OpenAI, Anthropic, or a custom OpenAI-compatible endpoint.

SSO is required for user authentication, with OIDC support for providers including Keycloak, Microsoft Entra ID, Okta, GitHub, GitLab, Google, and Ping Identity. The SaaS setup guide says the first AI Adoption Report baseline is available within seven days of installation. For questions the documentation does not resolve, support is available by email.

The key integration caveat is source control: CodeTogether's current AI documentation supports GitHub as a source-control connector, not Bitbucket, GitLab, or Azure DevOps. Its integrations page names Jira and several source-control and editor products, but that does not remove the stated connector limitation.

Pricing

CodeTogether is freemium, with a free plan and a free trial. The paid tiers are priced per developer per month, so team cost scales with headcount; both list a 10% discount for annual licenses.

• Starter Plan: $10 a month per developer, for up to 25 developers. It is the entry paid option for smaller teams, but its team-size cap rules it out for larger deployments.
• Business Plan: $49 a month per developer, with unlimited team size. It fits teams that need to grow beyond Starter's 25-developer ceiling, at a substantially higher per-developer price.
• Enterprise Plan: custom pricing, with custom quotes available for larger teams, on-premises deployment, and SSO support. It is the relevant tier for organizations needing those enterprise arrangements.

The free plan's price and specific limits are not given, so buyers should not assume it covers a particular team size or workload. The listed trial has no stated duration.

Platforms

CodeTogether is available as an extension, on Linux, macOS, and Windows, through the web, and as a self-hosted deployment. That range supports teams working across desktop editors and browser participation, as well as organizations that need an on-premises option.

Who it's for

CodeTogether is a strong fit for engineering organizations that want to relate AI spend and developer time to pull-request value, especially those working across multiple supported IDEs or requiring private-network deployment. It is a weaker fit for teams that need GitLab, Bitbucket, or Azure DevOps as the AI source-control connector, or that want voice and screen sharing in the collaboration tool itself.

Pros and cons

• Pro: It ties AI spend and human time to delivered pull requests, giving managers a value-per-dollar view rather than usage figures alone.
• Pro: Cross-IDE collaboration and browser joining accommodate teams that do not share one editor.
• Pro: SaaS and on-premises deployment options, plus stated code-handling safeguards, give organizations different ways to address data boundaries.
• Con: GitHub is the only supported source-control connector in the AI documentation, limiting teams built around other providers.
• Con: There is no voice audio or screen sharing, so live collaboration lacks those meeting capabilities.
• Con: Starter caps at 25 developers, while Business costs $49 a month per developer, making larger teams face a sizable per-seat step up.

Alternatives

Pair Programming Software is a useful starting point for comparing tools in the broader category.

• Drovio is worth considering when screen sharing and calls matter more than AI work measurement; its free Basic plan allows up to two participants and limits remote control to 30 minutes per day.
• Gobby is a free alternative for readers looking for software available on Linux, macOS, Windows, or self-hosted deployments.
• Pop is an option for unlimited meetings without time limits on its free Pro plan while in beta.
• Zed may suit an individual seeking a free code-editor plan with 2,000 accepted edit predictions and external-agent or API-key support.
• HackerRank is a different choice for interview and screening workflows; its Starter plan is $165 a month, billed $1,990 annually, for one user and 120 attempts per year.
• Saros is a free, open-source alternative with no paid plans.
• Duckly may fit occasional sessions: its free plan supports two simultaneous participants and has no video calls.
• Visual Studio is the more direct choice for readers seeking a Windows IDE, with a free Community plan for students, open-source, and individual developers.

Verdict

Choose CodeTogether if your organization needs to understand AI-development costs and value at the pull-request level while enabling cross-IDE collaboration. The private-network option and range of supported editors strengthen its case for larger engineering teams. Look elsewhere if your source-control workflow depends on anything beyond GitHub, or if voice and screen sharing are essential to your shared sessions.