Kandev is an open source AI-powered development environment designed to orchestrate multiple coding agents in parallel. Rather than acting as a single AI coding assistant, Kandev provides a Kanban-style workspace where developers can assign tasks to different AI agents, review their work, manage Git worktrees, and create pull requests from a unified interface. It supports multiple AI providers, is self-hostable, and runs entirely without telemetry, making it an excellent choice for teams and developers who want greater control over AI-assisted software development.
As AI coding assistants become more capable, developers often want to run several agents simultaneously. Managing multiple terminal windows quickly becomes difficult, especially when reviewing changes, tracking conversations, and merging code.
Kandev solves this problem by providing a visual development environment with Kanban and pipeline views. Each task can run in its own Git worktree, allowing several AI agents to work independently without interfering with one another. Developers can then review generated code, compare changes, and approve pull requests from a single workspace.
Key Features Kandev
Multi-Agent Orchestration
Kandev can launch multiple AI coding agents simultaneously, allowing different tasks to be completed in parallel. This significantly improves productivity when working on larger projects.
Kanban Workflow
The built-in Kanban board helps organize development tasks, track progress, and visualize AI-assisted workflows in a familiar project management interface.
Integrated Code Review
Instead of switching between terminal windows, Kandev provides a unified workspace with:
File editor
Git diff viewer
File tree
Terminal
Browser preview
Agent conversations
This makes reviewing AI-generated code much more efficient.
Multiple AI Provider Support
Kandev supports numerous AI coding agents through the Agent Client Protocol (ACP), including:
Claude Code
OpenAI Codex
GitHub Copilot
Gemini CLI
Amp
Auggie
OpenCode
Developers are free to choose the provider that best fits their workflow.
Flexible Execution
Agents can run in several environments:
Local processes
Docker containers
Remote SSH servers
Cloud executors
This flexibility allows teams to scale workloads without relying solely on local hardware.
Git Worktree Management
Each task runs in its own Git worktree, reducing merge conflicts and allowing multiple features or bug fixes to be developed simultaneously.
Self-Hosted and Privacy Focused
Kandev is fully self-hostable and collects no telemetry. Organizations that handle sensitive source code can keep their development environment entirely within their own infrastructure.
Cross-Platform Support
The application runs on Windows, macOS, and Linux. It can be installed using Homebrew, NPX, npm, or by building directly from source.
Download Kandev v0.95.0 - Software Mirrors |
|---|
Kandev v0.95.0 for Windowskandev-windows-x64.zip | 149.32 MB kandev-windows-x64.tar.gz | 149.33 MB kandev-desktop-windows-x64-Kandev_0.95.0_x64-setup.nsis.zip | 102 MB kandev-desktop-windows-x64-Kandev_0.95.0_x64-setup.exe | 102 MB |
Kandev v0.95.0 for macOSkandev-macos-x64.tar.gz | 148.26 MB kandev-macos-arm64.tar.gz | 143.85 MB kandev-desktop-macos-x64-Kandev_0.95.0_x64.dmg | 153.58 MB kandev-desktop-macos-x64-Kandev.app.tar.gz | 156.92 MB kandev-desktop-macos-arm64-Kandev_0.95.0_aarch64.dmg | 149.17 MB kandev-desktop-macos-arm64-Kandev.app.tar.gz | 152.75 MB |
Kandev v0.95.0 for Linuxkandev-linux-x64.tar.gz | 148.46 MB kandev-linux-arm64.tar.gz | 142.46 MB kandev-desktop-linux-x64-Kandev_0.95.0_amd64.deb | 158.1 MB kandev-desktop-linux-x64-Kandev_0.95.0_amd64.AppImage.tar.gz | 209.15 MB kandev-desktop-linux-x64-Kandev_0.95.0_amd64.AppImage | 212.72 MB kandev-desktop-linux-x64-Kandev-0.95.0-1.x86_64.rpm | 158.11 MB kandev-desktop-linux-arm64-Kandev_0.95.0_arm64.deb | 152.68 MB kandev-desktop-linux-arm64-Kandev_0.95.0_aarch64.AppImage.tar.gz | 199.71 MB kandev-desktop-linux-arm64-Kandev_0.95.0_aarch64.AppImage | 203.19 MB kandev-desktop-linux-arm64-Kandev-0.95.0-1.aarch64.rpm | 152.69 MB |
Kandev v0.95.0 Source Code
Kandev v0.95.0 Source code (zip)
Kandev v0.95.0 Source code (tar.gz)
|
Kandev v0.95.0 Release Notes: [!WARNING]
macOS, Windows desktop installers in this release are unsigned development builds.
They may require manual OS security bypasses and are not trusted downloads.
>
If macOS blocks Kandev after you verify the checksum, run:
>
bash
xattr -d com.apple.quarantine /Applications/Kandev.app
0.95.0 - 2026-09-18
Features
- implement taskless runs and Office mode repairs (#3777)
- add webhook alert dedup, filters, and repo selector (#3719) by @nova28
- report routine schedule state and harden coordinator install (#3763) by @nova28
- let a scheduled coordinator agent see the board and flag blockers (#3723) by @nova28
- surface unanswerable questions in a new History tab (#3679) by @nova28
- make session capacity opt-in (#3790)
- add workflow import profile selection (#3771)
- make task actions keyboard accessible (#3767)
- add smooth chat animations and scrolling (#3758)
- add adaptive Kanban scrolling cues (#3739)
- show configured fallback in profile rows (#3754) by @Fclem
- show CLI provider icons in composer model picker (#3760)
- add prompt-history plugin to catalog (#3757) by @Fclem
- protect agent plan edits and recover revisions (#3745)
- scope sidebar task views by workspace (#3731)
- compute review-round entry number server-side (#3266) by @nova28
- preview workflow moves before execution (#3677)
- register Redmine marketplace entry (#3684) by @yattdev
- add OpenAI-compatible provider primitive for coding agents (#3233) by @samueltorres-io
- reorganize storage settings and improve retention UX (#3695)
- port agent_conversation Host contract from the maintained fork (#3672) by @yattdev
- add generic approval ledger and receipts (#3238) by @yattdev
- add a Failed tab to the Needs-you Inbox (#3683) by @nova28
- improve PR walkthrough impact summaries (#3706)
- add provider-neutral change request tools (#3671)
- add signed webhook automation adapters (#3692) by @meanderix
- let new alert sources share storage, not a table per integration (#3694) by @nova28
- add instance-wide agent session concurrency ceiling (#3687) by @nova28
- add prompt history plugin Host prerequisites (#3588) by @Fclem
- identify symlinks in changes and files (#3675)
- reveal collapsed sidebar on configurable hover (#3678)
- add optional tool payload retention (#3668)
- add explicit Docker workload recipe (#3648)
- compact task prompt chips (#3665)
- handle task change links (#3506) by @yattdev
- make new task auto-focus optional (#3653)
- restore Quick Chat tabs and expose agent goals (#3645)
- add canvas marketplace distribution and sharing (#3603)
- explain divergent contribution histories (#3627)
- group homepage view settings (#3658)
- add workspace Needs-you Inbox for pending agent questions (#3641) by @nova28
- add priority sort and filter to the kanban board (#3344) by @nova28
- bound run history growth with a scheduled retention sweep (#3566) by @nova28
- keep worktree and local agents running across a backend restart (#3467) by @nova28
- open canvas setup dialog directly (#3587)
- allow switching a task's executor before materialization (#3559) by @nova28
- add grid layouts and auto-hide composers (#3626)
- add a workspace-wide pause to halt autonomous agent launches (#3536) by @nova28
- drag-and-drop reorder kanban task cards within a column (#3623) by @nova28
- add temporary storage visibility and cleanup (#3633)
- show workflow step progress (#3638)
- make the unattended loop legible end-to-end (#3613) by @nova28
- push to a caller-named remote and branch, not just origin (#3558) by @nova28
- detect and tear down dead SSH session transports (#3581) by @nova28
- bring the pipeline task row to parity with the Kanban card (#3284) by @nova28
- show task attachments during workspace preparation (#3594)
- use native mobile confirmation sheets (#3621)
- improve mobile dashboard parity (#3614)
- save repository set base branches (#3269)
- group task and card menu actions (#3599)
- one-shot step-entry move overrides (#3389) by @luancm
- add task actions to Threads view (#3585)
- explain workflow approval attention (#3584)
- allow Threads as the default Home view (#3576)
- make plan comments task-scoped (#3332)
- show database footprint in storage analysis (#3575)
- standardize responsive control sizing (#3563)
- add task actions menu to preview panel and detail top bar (#3395) by @nova28
- mark unstartable tasks with orphaned workspaces on the board (#3475) by @nova28
- add explicit session targeting (#3557)
- fail closed on budget admission for unattended runs (#3520) by @nova28
- let operators return a paused or stopped Office agent to idle (#3534) by @nova28
- fence queued message editing against concurrent sends (#3480) by @Fclem
- expose agent-accessible settings (#3543)
- improve Kubernetes worker lifecycle visibility (#3553)
- queue prompts during session resume (#3554)
- migrate workflow decisions to task-bound CLI (#3484)
- add Apprise availability rescan (#3551)
Bug Fixes
- preserve step prompt across async agent start failure (#3792) by @nova28
- stop sentry and gitlab from leaking other workspaces' issue watches (#3681) by @nova28
- retry a transient agent failure before it auto-pauses the agent (#3725) by @nova28
- stabilize workflow move preview invalidation (#3787)
- repair chat markdown section separators (#3785)
- recover stale workflow completion signals (#3772) (#3788)
- restore session-open conversation recovery (#3789)
- preserve resumed agents after accepted turn cancellation (#3791)
- preserve workflow prompts after async startup failure (#3753) (#3768)
- preserve repository discovery results and recovery (#3766)
- prevent ceiling replay lock cycles (#3779)
- restore prompt readiness after remote reconnect (#3774)
- break ceiling replay boot deadlock (#3776)
- coalesce streamed agent plan messages (#3764)
- retry transient Cursor RetriableError failures (#3765)
- coordinate maintenance health and compaction status recovery (#3770)
- align task summary semantic equality (#3769)
- preserve queued session ownership and visibility (#3755)
- unify executor profile editor routes (#3740) (#3762)
- routine create/save persists every field and can arm a cron schedule (#3759) by @nova28
- stop scrambling and truncating dynamic-agent launch prompts (#3741) by @nova28
- stop mark fixed from silently discarding a newer pause reason (#3734) by @nova28
- show archive progress before removal (#3730)
- preserve file-tree rows after hidden panel restore (#3744)
- improve condensed workflow hover interaction (#3746)
- warn when auto-start is skipped for unresolved dependencies (#3724) by @benjamincham
- name unknown tool arguments and state task binding (#3733) by @benjamincham
- approve Claude ACP Kandev MCP permissions (#3729) (#3738)
- compact fully integrated managed branches (#3158) by @yattdev
- quiet shutdown logs, portable MCP tool schemas, and Copilot model UX (#3726)
- retract output from abandoned provider retries (#3728)
- restore Nest under candidates (#3727)
- remove mobile task dragging (#3701)
- keep transient retry banners to one notice (#3704)
- don't fire paused or archived routines on cron, manual run, or webhook (#3535) by @nova28
- pair provider_scope with provider_repo_id on repository writes (#3699) by @ClemDNL
- restore cancel controls during active turns (#3705)
- protect queued FIFO dispatch admission (#3702)
- enforce exact profile model identity (#3473) by @yattdev
- recover sessions after lost Pods (#3703)
- stop the needs-you inbox from flashing Loading on background refresh (#3685) by @nova28
- restore single-user MCP PR linking (#3698)
- forward TaskID in the Docker create-instance request (#3612) by @StoyKK
- discover locally installed agents across launch modes (#3696)
- make retained PVC preparation robust (#3697)
- archive a task whose canonical environment is gone (#3659) by @nova28
- make review finding status updates idempotent on resolved_at (#3670) by @nova28
- rename the Inbox feature toggle label to match the sidebar (#3673) by @nova28
- support direct agent profile invocation (#2870) by @yattdev
- persist right-panel visibility across layouts (#3657) (#3661)
- retain session failures across recovery surfaces (#3669)
- make Docker worker lifecycle reliable (#3676)
- expose accepted workspace mode values (#3667)
- harden task archive lifecycle recovery (#3651) by @Fclem
- widen workflow hover disclosure (#3664)
- scan aliased and dot-imported os environment reads (#3505) by @yattdev
- preserve PR watch branch identities (#3649)
- allow scrolling through long clarification footers (#3655)
- preserve plan comment drafts on foreground refresh (#3652)
- bound Git subprocess execution (#3646)
- harden startup recovery timeout and scrolling (#3654)
- make routine catch-up enqueue a summarized wake (#3538) by @nova28
- stop provider failure diagnostics from being lost or misattributed mid-stream (#3479) by @nova28
- keep budget alerts honest when a policy changes mid-check (#3532) by @nova28
- normalize workflow move options styling (#3644)
- harden canvas review and runtime bootstrap (#3628)
- recover slow analytics and workspace reads (#3639)
- stop duplicate and missed children-completed run dispatches (#3620) by @nova28
- keep an agent's session alive while it's still working the task (#3556) by @nova28
- recover plan comments without blocking empty chat (#3616)
- bridge executor HTTPS Git access through host gh (#3072) (#3635)
- recover delayed session entry (#3631)
- fence cancelled session recovery attempts (#3636)
- preserve task brief on first chat message (#3615) (#3632)
- make repeat task reassignment work again after 24 hours (#3533) by @nova28
- reap orphaned processes left running after workspace cleanup (#3619) by @nova28
- stop double-dispatching step-entry on_enter actions (#3288) by @nova28
- tear down orphaned agent processes when a session never starts (#3394) by @nova28
- reuse documentation requirement lookups (#3640)
- stop plugin pending-interaction API from picking the wrong current turn (#3618) by @nova28
- compact multiple workflow swimlane heights (#3604)
- correct cron DOM/DOW OR semantics, DST fire policy, and unsatisfiable expressions (#3514) by @nova28
- stop office tasks reaching Done while stuck on an earlier workflow step (#3392) by @nova28
- keep large database startup alive (#3624)
- keep repository actions available (#3593)
- restore completed task workspace access (#3577)
- scope PR auto-fix outcome reporting (#3602)
- remove archived chats immediately (#3625)
- honor same-profile workflow session policy (#3605)
- retry on_agent_error recovery after a failed commit instead of giving up (#3441) by @nova28
- fence session recovery and preserve fallback outcomes (#3610)
- separate task creation by workspace mode (#3607)
- restore pull request discovery health (#3609)
- stop over-budget policies from spamming duplicate alerts (#3287) by @nova28
- resolve launch repositories from the task attachment set (#3582) by @nova28
- honor selected remote PR branches (#3591)
- condition discard consent on local changes (#3590)
- stabilize canvas runtime and permission review (#3600)
- require explicit handoff context selection (#3601)
- preserve content-only checkouts (#3137) by @yattdev
- scope legacy empty-branch environment fallback per repository (#3595) by @WaleWangPW
- keep iOS composer suggestions visible above the keyboard (#3596)
- stop orphaned agentctl processes from piling up on SSH runners (#3527) by @nova28
- prevent selected task removal navigation flicker (#3578)
- resolve worktree dir by sanitized repo name on launch admission (#3580) by @JnManso
- stop agent heartbeat cooldowns from being silently ignored (#3512) by @nova28
- stop parent tasks from stalling when a reopened child finishes again (#3290) by @nova28
- stop provider fallback from leaking secrets and raw continuation text (#3393) by @nova28
- prevent duplicate or stranded launches when tasks auto-start on create (#3529) by @nova28
- improve mobile threads and shared listing headers (#3570)
- block workflow proceed during clarification (#3574)
- include npm global bin in service path (#3571) (#3573)
- stop different tasks' queued runs from silently merging (#3517) by @nova28
- blur mobile menu sheet backgrounds (#3569)
- detach inherited environment on executor mismatch (#3565) by @WaleWangPW
- restore completed task conversation follow-ups (#3564)
- preserve signal-gated workflow proceed action (#3572)
- don't leave task views stale after an approval-gate redirect (#3492) by @nova28
- don't flag healthy sessions as orphaned after upgrading (#3524) by @nova28
- stop a task status from reverting when an older update finishes late (#3528) by @nova28
- stop reporting a dropped backend response as success (#3562) by @nova28
- stop a parent-lookup DB failure from surfacing as a 403 permission denial (#3491) by @nova28
- respect dark mode for initial toast notifications (#3568)
- stop the stuck-parent wake sweep from failing every tick on Postgres (#3525) by @nova28
- show already-attached skills as checked in the agent skills tab (#3200) by @nova28
- stop runner lookups from failing on Postgres (#3289) by @nova28
- retain eligible providers during refresh (#3498) by @yattdev
- group sidebar trees by effective state (#3561)
- make the board search box match linked PR/MR numbers (#3391) by @nova28
- keep the chat composer focused after sending a message (#3537) by @nova28
- don't collapse a genuine clarification/permission wait to REVIEW (#3549) by @nir-hossted
- make Pi MCP eager and pin adapter (#3546) by @edan-binshtok
Performance
- pause hidden status motion (#3589)
Documentation
- decline the internal/reposync extraction and reconcile the Office config-sync design (#3717) by @nova28
- simplify GitHub credential guidance (#3630)
- improve planning assumption checks (#3629)
- replace tracked document catalogs (#3482)
- clarify single-binary distribution (#3592)
- remove duplicated runtime default claims (#3608)
- note that agent_profile_id cannot set an Office task assignee (#3526) by @nova28
- correct the reason a blocking MCP question survives 5 minutes (#3523) by @nova28
- document the ACP session-transition lock so new session paths do not bypass it (#3396) by @nova28
|
Performance and User Experience
Kandev offers a modern interface that combines project management with AI-assisted development. The Kanban workflow feels natural for teams already using agile methodologies, while the integrated review tools reduce context switching between editors, terminals, and Git clients.
The application performs well even when managing multiple concurrent AI sessions, thanks to its orchestrator architecture and support for isolated execution environments. Since it is still a relatively young project, some features continue to evolve rapidly, but active development and frequent updates indicate strong momentum.
Pros
Free and open source.
Supports multiple AI coding agents.
Parallel task execution.
Built-in Kanban workflow.
Excellent integrated code review tools.
Self-hostable with no telemetry.
Flexible execution using local, Docker, SSH, or cloud environments.
Active development.
Cons
Still a relatively new project.
Best suited for experienced developers.
Requires compatible AI agent tools and API access.
Advanced workflows may require some initial configuration.
Who Should Use Kandev?
Kandev is ideal for software developers, engineering teams, AI-assisted coding enthusiasts, and DevOps professionals who frequently work with multiple AI coding agents.
It is especially valuable for teams that want to coordinate parallel AI workflows, maintain human review before merging code, and avoid vendor lock-in by supporting multiple AI providers.
Final Verdict
Kandev introduces a fresh approach to AI-assisted software development by treating AI agents as collaborative team members rather than standalone chatbots. Its combination of multi-agent orchestration, Kanban project management, integrated code review, and self-hosted architecture makes it one of the most promising open source AI development environments currently available.
Although the project is still evolving, its rapid development pace, privacy-focused design, and flexible architecture make it an excellent choice for developers looking to scale their AI-assisted workflows.
Post a Comment/Report Broken Link: