Google Jules

Deep GitHub integration covers issue labels, branches, pull requests and reading and responding to PR comments, MCP servers connect Linear, Stitch, Neon, Tinybird, Context7 and Supabase, a Render integration detects failed builds and pushes fixes, and the REST API wires Jules into Slack, Linear, Jira and CI pipelines including GitHub Actions; the agent runs shell commands, installs dependencies and executes builds inside its VM with repository level environment variables.

Sourcejules.google/docs/changelog and jules.google/docs/integrations, retrievedread 2026-08-30

A task runs as a multi step plan the agent produces before writing code, with an Interactive Plan mode where it asks clarifying questions and refines the approach collaboratively, a Critic agent performing adversarial review of every proposed change during generation, a Planning Critic reviewing auto approved plans and reported to cut task failure rates by 9.5 percent, a CI Fixer loop that detects failures, fixes, commits and resubmits, and up to five parallel sessions on one prompt via the CLI --parallel flag.

Sourcejules.google/docs/changelog 2026-01-26-1, 2026-02-19 and 2025-11-10, retrievedread 2026-08-30

Each task clones the repository into its VM and reads it directly, with AGENTS.md conventions, a file selector to scope context, image upload and web search for current documentation. Context is assembled per task, and no persistent index, graph or embeddings layer over the customer's code is documented.

Sourcejules.google/docs/changelog 2025-09-29, 2025-08-08 and docs/environmentread 2026-09-29

The agent writes a step by step plan the user can review, edit or reject before any code is written, with an Interactive Plan mode asking clarifying questions first; work lands on a branch and the user chooses when to publish or open a pull request, a Critic agent adversarially reviews each proposed change during generation and a Planning Critic reviews plans that are auto approved, tasks can be paused, resumed or deleted mid run, and commit authorship is configurable so changes can be attributed to Jules, co-authored, or the user.

Sourcejules.google/docs/changelog 2025-08-082, 2026-01-26-1 and 2026-02-19, retrievedread 2026-08-30

Jules documents task isolation in a cloud VM, states it does not train on private repositories and tells users they are responsible for the code they run, but no access surface (SSO, roles, organization administration) and no attestation or trust page is documented, and paid plans are sold to individual Google accounts only.

Sourcejules.google/docs/faq and docs/usage-limitsread 2026-09-29

Each session carries a collection of activities representing the steps the agent took, described by the vendor as immutable, following an event source pattern and usable to reconstruct the session, retrievable through the REST API with a createTime range cursor; the full changeset is downloadable in git patch format, the critic agent's step by step evaluation is visible in the UI in real time, and the plan is shown before execution.

Sourcejules.google/docs/changelog 2026-01-26-4 and 2025-09-03, retrievedread 2026-08-30

Repository Memory saves the user's preferences, nudges and corrections during a task and references them on later tasks in that repository, toggled on or off per repository under Knowledge. Memory states no lifetime and no way to review or delete entries, and environment snapshots and variables are setup state rather than memory.

Sourcejules.google/docs/changelog 2025-09-30read 2026-09-29

Every task runs in a Google managed cloud VM whose toolchain the customer shapes with a setup script, snapshots and environment variables; that configures the environment rather than where it runs, and no region choice, customer VPC, on premises or tenant option is documented, so a buyer with a different residency requirement has no other answer.

Sourcejules.google/docs/environment and docs/faqread 2026-09-29

Sample prompts ship on the home page as one click starting points, suggested tasks let the agent propose work it has found in the repository, and MCP integrations are offered from a vendor curated list of vetted servers. There is no library of prebuilt agents, installable skills or shareable task templates.

Sourcejules.google/docs/changelog 2025-09-02, 2025-12-10 and 2026-02-02, retrievedread 2026-08-30

Tasks start from a jules label on a GitHub issue, the web app, the Jules Tools CLI, a REST API for CI/CD and chat integrations, scheduled recurring tasks with edit, pause and resume, proactive suggested tasks where Jules scans a repository and surfaces work, pull request comments which the agent reads and acts on, and automatic triggers on CI failures in GitHub Actions and failed Render builds.

Sourcejules.google/docs/changelog scheduled tasks, suggested tasks and CI fixer entries, retrievedread 2026-08-30

Every model Jules runs is Google's Gemini: Gemini 3 Flash is the base model on all tiers and Gemini 3.1 Pro the default for Pro subscribers, which is a tier choice within one maker, and no other maker or bring your own model path is documented.

Sourcejules.google/docs/changelog 2026-03-09 and 2026-01-30, docs/usage-limitsread 2026-09-29

MCP server support launched February 2026 with a vetted initial set covering Linear, Stitch, Neon, Tinybird, Context7 and Supabase, configured by API key in settings and invoked when the agent detects a tool need; a REST API at jules.googleapis.com/v1alpha exposes Sources, Sessions and Activities with API key auth, a Jules Tools CLI installs via npm as @google/jules with parallel execution and local patch application, a Gemini CLI extension exposes Jules as a slash command, and a Render integration is documented.

Sourcejules.google/docs/changelog 2026-02-02 and 2025-10-03, jules.google/docs/api/reference/authentication, retrievedread 2026-08-30

The agent runs the project's existing test suite and iterates when tests fail, writes tests on its own, and the CI Fixer detects a failed GitHub Actions check, works through a fix and resubmits the pull request in a loop; a Critic agent adversarially reviews proposed changes and a Planning Critic reviews plans, both internal quality mechanisms. These verify the customer's code and the agent's own output rather than providing a customer facing harness to evaluate or regression test agent behavior.

Sourcejules.google/docs/changelog 2026-02-19 and 2025-08-083, retrievedread 2026-08-30

The default Jules base image includes Playwright for front end testing, and the agent renders a running web application and returns a screenshot of the front end alongside its code changes when asked to verify its work; the Gemini 3 Pro release notes cite visual verification of web app outcomes with higher precision, and Jules can also proactively search the web for documentation and code examples during a task.

Sourcejules.google/docs/changelog 2025-08-07 and 2025-11-19, retrievedread 2026-08-30