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 v0.94.0 - Software Mirrors |
|---|
Kandev v0.94.0 for Windowskandev-windows-x64.zip | 145.17 MB kandev-windows-x64.tar.gz | 145.18 MB kandev-desktop-windows-x64-Kandev_0.94.0_x64-setup.nsis.zip | 99.02 MB kandev-desktop-windows-x64-Kandev_0.94.0_x64-setup.exe | 99.02 MB |
Kandev v0.94.0 for macOSkandev-macos-x64.tar.gz | 144.15 MB kandev-macos-arm64.tar.gz | 139.97 MB kandev-desktop-macos-x64-Kandev_0.94.0_x64.dmg | 149.52 MB kandev-desktop-macos-x64-Kandev.app.tar.gz | 152.76 MB kandev-desktop-macos-arm64-Kandev_0.94.0_aarch64.dmg | 145.32 MB kandev-desktop-macos-arm64-Kandev.app.tar.gz | 148.78 MB |
Kandev v0.94.0 for Linuxkandev-linux-x64.tar.gz | 144.37 MB kandev-linux-arm64.tar.gz | 138.64 MB kandev-desktop-linux-x64-Kandev_0.94.0_amd64.deb | 153.94 MB kandev-desktop-linux-x64-Kandev_0.94.0_amd64.AppImage.tar.gz | 205.49 MB kandev-desktop-linux-x64-Kandev_0.94.0_amd64.AppImage | 209.01 MB kandev-desktop-linux-x64-Kandev-0.94.0-1.x86_64.rpm | 153.94 MB kandev-desktop-linux-arm64-Kandev_0.94.0_arm64.deb | 148.77 MB kandev-desktop-linux-arm64-Kandev_0.94.0_aarch64.AppImage.tar.gz | 196.34 MB kandev-desktop-linux-arm64-Kandev_0.94.0_aarch64.AppImage | 199.8 MB kandev-desktop-linux-arm64-Kandev-0.94.0-1.aarch64.rpm | 148.77 MB |
Kandev v0.94.0 Source Code
Kandev v0.94.0 Source code (zip)
Kandev v0.94.0 Source code (tar.gz)
|
Kandev v0.94.0 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
0.94.0 - 2026-09-09
Features
- improve MCP discovery and canvas authoring prompts (#3522)
- move automation toggles to session scope (#3465) by @Fclem
- show a distinct icon for tasks parked on background work (#3402) by @nova28
- enforce required store parity (#3459)
- resolve unique advertised model variations (#3472)
- confirm saved view deletion (#3461)
- activate reviewed runtime defaults on startup (#3471)
- simplify launch-step destination (#3466)
- serve native HTML previews (#3375)
- preview task launch destination and prompt (#3425)
- add automatic task color rules (#3180)
- add progressive storage analysis (#3429)
- add DeepSeek API Balance to plugin registry (#3452) by @Fclem
- add append mode to update_task_plan_kandev (#3448) by @nova28
- add consent-aware repository discovery
- per-profile opt-in for unprivileged user namespaces by @yattdev
- expand task fields for integrations by @yattdev
- add chat-submit-decoration plugin slot over the send button (#2990)
- add plugin-backed canvases (#3111)
- organization units, tenancy, roles and scopes for team access (#2919)
- add task manager to marketplace (#3406)
- show and set task priority on the kanban board (#3333) by @nova28
- move shortcuts into plugin settings (#3384)
- resolve session hostnames (#2741) by @Fclem
- show and move a task's workflow step from the kanban preview (#3329) by @nova28
- add Goose ACP agent support (#3313) by @jelloeater-agent
- support right-click session actions in the kanban preview panel (#3321) by @nova28
- split system data and storage pages (#3348)
- add Google Antigravity as a selectable coding agent (#3324) by @nova28
- sync agents, projects, skills, and routines from GitHub or GitLab (#3304) by @nova28
- let the dependency picker find tasks by PR or MR number (#3311) by @nova28
- add a Plan tab to the kanban preview panel (#3314) by @nova28
- distinguish archived command-panel task results (#3298)
- add Threads view for live agent conversations (#3112) by @jcoatelen-ledger
- add manual file upload and download to the Files panel (#3292) by @Corey-Fogg
- reject task plan saves over 256 KiB instead of growing unbounded (#3297) by @nova28
- add workflow profile session lifecycle policies (#3225)
- support dual-era MCP protocol (#3164)
- surface stalled Office tasks without reclaiming them (#3279) by @nova28
- filter issues by milestone (#3278) by @nova28
- add Kubernetes executor (#3261)
- show char count, delta, and workflow step in plan history (#3265) by @nova28
Bug Fixes
- restore MCP status after task switches (#3547) (#3552)
- preserve queued turn state during workflow transitions (#3542)
- recover cleanup preparation for missing worktrees (#3541)
- expand saved prompts on workflow-free launches (#3539)
- stop multi-agent reviews from deadlocking by default (#3515) by @nova28
- parse run-e2e.sh options after a leading -- and bound the docker info probe (#3518) by @nova28
- reject empty slug on skill update (#3509) by @nova28
- stop routine tasks from getting office skills they can't run (#3502) by @nova28
- protect referenced secrets from deletion (#3500) (#3503)
- guard archived session recovery and preserve launch errors (#3519)
- remove host model warnings from profile selectors (#3521)
- stop agents showing idle while a successor run works (#3516) by @nova28
- name the children in the children-completed wake prompt (#3510) by @nova28
- retry SSH agentctl handshake rejection instead of failing the launch (#3511) by @nova28
- stop discarding step_complete's handoff before the next step reads it (#3446) by @nova28
- stop messages sent during a step move from vanishing (#3494) by @nova28
- retract empty-turn notice once real agent output lands (#3499) by @nova28
- stop DST fall-back from firing automations twice (#3490) by @nova28
- stop routines from bricking after their first fire (#3488) by @nova28
- tell the CEO which agent failed, not "Error: unknown" (#3485) by @nova28
- prevent plan editor crash during task switches (#3493)
- recover managed runtime capability probes (#3483)
- address Dependabot vulnerabilities (#3450)
- recognize visible Dockview chat panels (#3481)
- grant record_step_decision capability to decision-seat holders (#3463) by @nova28
- preserve pending git action ownership (#3373) by @yattdev
- use trusted preview deploy command (#3456) by @yattdev
- hide inactive repository providers (#3470)
- preserve scroll position across task switches (#3474)
- stabilize archive confirmation surface (#3460)
- stop Mark fixed from leaving an auto-paused agent stuck (#3464) by @nova28
- stabilize failing main e2e runs (#3462)
- recover failed task launches safely (#3437)
- retry PR auto-fix turns without outcomes (#3458)
- stop restarts from leaving the API stuck in "starting" (#3445) by @nova28
- stop a manual route retry from deadlocking against a context reset (#3443) by @nova28
- stop workflow triggers from re-running actions right after a restart (#3447) by @nova28
- stop reviewer and approver comments showing under the assignee's name (#3444) by @nova28
- preserve PR status colors across reload (#3454)
- bound clarification response submissions (#3439)
- refine queued message row controls (#3449) by @Fclem
- stabilize flaky e2e workflows (#3453)
- protect dirty worktree deletion (#3426)
- preserve Kanban column width while dragging (#3345) (#3423)
- suppress invalid task timestamps (#3434)
- keep PR walkthrough links canonical (#3438)
- break approver-first tie when neither quorum seat sits at the current step (#3442) by @nova28
- stop Sprites prepare truncating git history and leaking the clone token (#3411) by @jmarbutt
- keep draft PR icons muted (#3424)
- keep new-file input focused after menu close (#3419)
- auto-fix GitHub PR merge conflicts (#3418)
- hide human assignee when auth is disabled (#3417)
- classify backend reload diagnostics (#3414)
- stop dynamic agent handoffs from leaking secrets and mangling text (#3407) by @nova28
- show agents as working while a run is in flight (#3401) by @nova28
- stop manual Retry from starting a second agent when a route already resumed (#3399) by @nova28
- read the stdin environment to EOF for bash on macOS (#3385) by @jmarbutt
- brace refspec variables for zsh login shells (#3383) by @jmarbutt
- harden prompt history and saved prompt aliases (#3369) by @Fclem
- contain mobile Kanban archive confirmation (#3430)
- preserve remote OpenCode auth providers (#3382) (#3421)
- preserve latest task-create workflow (#3413)
- load and scroll to unloaded prompts (#3412) by @Fclem
- preserve workflow move handoff prompts (#3409) (#3415)
- keep a usable agent selected when the executor changes (#3388) by @jmarbutt
- stop a stale session snapshot from reverting a model switch (#3379) by @JnManso
- preserve transcript position on task switch (#3405)
- stop a slow MCP server from stalling Claude for hours (#3398) by @nova28
- stabilize immutable runtime image resolution (#3397) by @yattdev
- align merge approval cleanup contract
- repair stale approval merge-state query
- restore PostgreSQL domain store parity (#3372)
- open registered repository links in file viewer (#3400)
- make passthrough composer touch-safe (#2809) (#3303)
- stop dynamic-route retries from double-launching the same prompt (#3360) by @nova28
- preserve detached workspace ownership (#3366) (#3374)
- recover dynamic agent routes stuck in starting with no owner (#3362) by @nova28
- use built-in token for runtime pin PRs (#3365)
- reject stale executor transition workspaces (#3354) by @WaleWangPW
- stop disabled dynamic routing from resuming pending retries (#3361) by @nova28
- backfill feeder pulls after task-tree lifecycle changes (#3376)
- make accordion menu the default (#3368)
- launch the dynamic fallback successor outside the agent.failed dispatch (#3356) by @iamapfelbaum
- prevent silent internal bridge request loss (#3364) (#3367)
- compact TUI MCP strategy selector (#3370)
- make kanban boards recover from agent failures too (#3315) by @nova28
- stop retries from hammering a failed credential after a restart (#3357) by @nova28
- resolve Windows resume share-access violation (#3353) by @JnManso
- keep change action spinners visible (#3339)
- detect backend restarts in open pages (#3346)
- quiet Claude passthrough by default (#3305) (#3317)
- classify OpenCode period usage limits so dynamic routing can advance (#3320) by @iamapfelbaum
- restore compact coarse-pointer file tree rows (#3326) by @Fclem
- keep model selector visible after model switches (#3349)
- recover pathless worktree failures (#3335) (#3351)
- recover SSH sessions after dead runtime (#3330) (#3350)
- stop superseded session worktrees claiming cutover slots (#3328) by @tito
- registering a reviewer stops duplicating the auto-picked seat (#3327) by @nova28
- stop workflow step transitions from leaking claude subprocesses (#3338) by @nova28
- fixes MCP issue where auto_approve is silently ignored in agent profile create/update. (#3331) by @meanderix
- stop clarification answers from timing out on large message tables (#3322) by @nova28
- retain PR commit provenance across refreshes (#3341)
- restore chat scroll on env-changing task switch (#3342)
- enable coarse-pointer terminal scrolling (#2808) (#3301)
- stop chat from saying no utility agent is configured after you save one (#3306) by @nova28
- deduplicate empty workflow prompts (#3299)
- stop the reject leg from waking the reviewer, not the assignee (#3294) by @nova28
- guard session resumption across task navigation (#3302)
- contain a panicking subscriber so it stops crashing event delivery (#3295) by @nova28
- stop a failing CEO agent from re-queuing itself forever (#3296) by @nova28
- stop parent tasks waking twice when children finish (#3271) by @nova28
- preserve qualified remote refs in watcher branches (#3268)
- preserve transcript scroll across session activation (#3291)
- cap the plan text pasted into every new agent session (#3280) by @nova28
- recognize reviewer decisions from agents who also hold the approver seat (#3270) by @nova28
- escalate reviewer and approver agent failures to the CEO (#3275) by @nova28
- fire budget alerts when reassigning a task pushes a project over budget (#3276) by @nova28
- materialize attachments sent into a generating turn (#3277) by @Corey-Fogg
- allow typing while sessions start (#3285)
- preserve plan comments across session switches (#3272)
- stop idle-skip from swallowing manual wakeups queued behind a cron run (#3281) by @nova28
- allow the stall watchdog to archive its own task (#3260) by @nova28
- stop failed answer submits from vanishing silently (#3263) by @nova28
- normalize skill slugs to canonical form on write and sync (#3257) by @nova28
Performance
- stream frontend diagnostic captures (#3540)
- isolate task surface rendering (#3226)
- isolate sidebar task tree renders (#3435)
- isolate kanban task update renders (#3436)
Refactoring
- remove legacy websocket request API (#3431)
- remove unused recreator (#3432)
- remove deprecated model fetcher cache (#3433)
- stop hand-syncing the dispatcher's active-state list with the SQL filter (#3274) by @nova28
Documentation
- correct stale unique-index precondition for office session identity (#3513) by @nova28
- mark shared access features experimental (#3428)
- add branded diagram design guidance (#3316)
- stop postgres and nats bugs from shipping green through verify (#3273) 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: