AgentVera features
Running coding agents side by side is just the start. These features connect them, manage limits and cost, keep the code in check and put the tools they need in the same window.
Work between agents
- Team view Turn your open agents into a team with a manager, roles and reporting lines, then watch it work live.
- FlowsPro Wire agents together on a canvas; when one finishes, its reply goes to the next.
- Parallel teamPro An agent splits a big job into parts; each runs in its own worktree in parallel and ends as one branch and one PR.
- Task linePro Queue tasks and let each pass through Analysis, Plan, Build and Review agents in order.
- Agent coordination Warnings when agents edit the same file, and a shared board for every agent group.
- Agent messaging Agents in a group ask each other questions, wait for answers and share a task list, even across different CLIs.
- Team timeline What a project’s or group’s agents did together, day by day and live.
- Night shiftPro Queue tasks in the evening and give each to an agent; agents work overnight and you get a morning report.
- Scheduled tasks Message an agent on a cron schedule; if it’s busy, wait, send or skip.
- Notifications and approvals Agents waiting for approval in one queue; desktop notifications, a Dock badge and Telegram.
- AutomationsPro An agent starts on its own when CI breaks, a PR gets a comment, an issue is labelled or a schedule fires.
- Merge queuePro Agent branches land one at a time, on the latest target and only after the check passes.
- Background agents and supervisor Slots and usage limits for background work, a supervisor for stuck agents, agents that open agents.
- Telegram control Start, manage and approve agents from your phone with the AgentVera bot; each agent gets its own topic.
Cost and limits
- Automatic account switchingPro Line up accounts for Claude, Codex and other agents; when one hits its limit, the agent moves to the next.
- Token savings Context profiles, shorter command output with RTK, live context size and auto-compact, on by default for every agent.
- Token budgets Token caps per agent, per turn and per day; a warning at 80%, and at the cap the agent stops but keeps its conversation.
- Automatic model Each task is rated simple, normal or hard, and the agent switches to that tier’s model.
- Smart tools Cuts repeat reads, full test runs and repeated commands before they reach the model.
- AI ROI analytics Agent success rate, tokens per task, follow-ups and commits that needed fixing later.
Code and quality
- Automatic code review When a turn ends, a second model reviews the agent’s code; findings link to file:line.
- Review inbox Every agent’s finished work in one list: changed files, risk notes, the Verify result and a one-click decision.
- AgentVera VerifyPro The acceptance criteria in your message are checked against evidence by an independent verifier at turn end.
- Auto approvalPro Approve Claude Code’s safe requests without asking; risky ones still come to you.
- File ownership Give an agent its own files and folders; while it runs, other agents stay out.
- Checkpoints and session replay A snapshot of every turn; undo one in a click and replay a session step by step.
- Project map Agents find the relevant files, their users and tests without reading file after file.
- AgentVera Brain A project memory that pulls decisions from docs, commits, PRs and agent conversations.
- Project tour and code map Get to know a project with a tour, see module dependencies and the files that break most.
- Git and GitHub Changes, graph history, worktrees, generated commit messages and GitHub pull requests.
- Conversation search Full-text search across every Claude and Codex conversation, including ones started outside AgentVera.
Delivery and deploys
- CI/CD center GitHub Actions, GitLab CI and Jenkins runs in one list; hand a broken run to an agent.
- Preview environments The dev server of the project and of every worktree on its own address, with visual diffs.
- Vercel Production and preview deployments, build logs, promote and roll back.
- DB Guard Migrations written by agents are scanned for risk at turn end, so data loss and locks show up early.
Tools
- MCP manager A catalog of 23 ready servers, MCP for every CLI agent, one sign-in for remote servers and servers agents ask for.
- Skill gallery Manage Claude Code and Codex skills, install scanned skills from the gallery, share your own.
- Database managerPro Tables, a SQL editor, ER diagrams and agent access for PostgreSQL, MySQL and SQLite.
- Send to agent Send terminal output, a line of code, a diff or logs to an agent in one keystroke.
- API client Find your project’s endpoints in the code, try them and send the response to an agent.
- Agent browser A separate-profile Chrome your agents drive with Playwright; watch its tabs live.
- Docker manager Containers, live logs, exec, images and compose, without opening Docker Desktop.
- Agents in Docker Run a Claude Code or Codex agent in a Docker container that only sees the project folder.
- SSH and SFTP Saved servers, an encrypted SSH vault, port forwarding, two-pane file transfer and agents on the server.
Team and account
- Team kanbanPro Team boards linked to your projects; hand a card to Claude, Codex or any agent and get the reply on the card.
- Mobile app The iOS and Android app: chat with the agents on your computer, approve requests, watch the terminal and changes.
- Review from your phone Your agents’ status, approval requests and finished work with diffs, in your phone’s browser.
- Account and organizations An AgentVera account, organizations, members, devices and billing, managed from the desktop.
Desktop experience
- Workspace A grid per project, ⌘K, remappable shortcuts, ports, light and dark themes, settings backup.
- Editor, terminal and specs Open and edit files, keep terminals under your agents, and see the specs agents write in one list.
- Group view Bring one group of agents to the front; watch their terminals or summary cards and act in bulk.
- Tasks and reminders A personal task list with quick add, reminders on desktop and Telegram, and one-click hand-off to an agent.
- Voice dictation Say your message; local Whisper turns it into text and the audio never leaves your computer.
Setup guides
- AgentVera Runner setup guideAgentVera Runner setup guide: install it on a Linux or macOS server with one command, pair it, and run Claude Code, Codex, Gemini CLI and opencode from the browser.
- AgentVera webhooks guideAgentVera webhooks guide: open kanban cards and hand Claude Code or Codex agents tasks automatically from GitHub, Sentry or your own system’s requests.
See the 18 supported coding agents
Bring your agents to one desk.
Start on the free plan. Download it, add a project folder and open your first agent.