Multiflow Docs

CLI Reference

Complete command reference for the Multiflow CLI and agent daemon.

The multiflow CLI connects your local machine to Multiflow. It handles authentication, workspace management, issue tracking, and runs the agent daemon that executes AI tasks locally.

Authentication

Browser Login

multiflow login

Opens your browser for OAuth authentication, creates a 90-day personal access token, and auto-configures your workspaces.

Token Login

multiflow login --token <mul_...>

Authenticate using a personal access token directly. Useful for headless environments. Pass --token= with an empty value to be prompted interactively (so the token never lands in shell history).

Check Status

multiflow auth status

Shows your current server, user, and token validity.

Logout

multiflow auth logout

Removes the stored authentication token.

Agent Daemon

The daemon is the local agent runtime. It detects available AI CLIs on your machine, registers them with the Multiflow server, and executes tasks when agents are assigned work.

Start

multiflow daemon start

By default, the daemon runs in the background and logs to ~/.multica/daemon.log.

To run in the foreground (useful for debugging):

multiflow daemon start --foreground

Stop

multiflow daemon stop

Status

multiflow daemon status
multiflow daemon status --output json

Shows PID, uptime, detected agents, and watched workspaces.

Logs

multiflow daemon logs              # Last 50 lines
multiflow daemon logs -f           # Follow (tail -f)
multiflow daemon logs -n 100       # Last 100 lines

Supported Agents

The daemon auto-detects these AI CLIs on your PATH:

CLICommandDescription
Claude CodeclaudeAnthropic's coding agent
AntigravityagyGoogle's coding agent
CodeBuddycodebuddyTencent's Claude Code-compatible coding agent
CodexcodexOpenAI's coding agent
CopilotcopilotGitHub's coding agent
Cursor Agentcursor-agentCursor coding agent
OpenCodeopencodeOpen-source coding agent
OpenClawopenclawOpen-source coding agent
HermeshermesNous Research coding agent
KimikimiMoonshot coding agent
Kiro CLIkiro-cliKiro ACP coding agent
PipiInflection coding agent
QoderqodercliAlibaba coding agent
Trae CLItraecliByteDance coding agent

You need at least one installed. The daemon registers each detected CLI as an available runtime.

How It Works

  1. On start, the daemon detects installed agent CLIs and registers a runtime for each agent in each watched workspace
  2. It polls the server at a configurable interval (default: 3s) for claimed tasks
  3. When a task arrives, it creates an isolated workspace directory, spawns the agent CLI, and streams results back
  4. Heartbeats are sent periodically (default: 15s) so the server knows the daemon is alive
  5. On shutdown, all runtimes are deregistered

Configuration

Daemon behavior is configured via flags or environment variables:

SettingFlagEnv VariableDefault
Poll interval--poll-intervalMULTICA_DAEMON_POLL_INTERVAL3s
Heartbeat interval--heartbeat-intervalMULTICA_DAEMON_HEARTBEAT_INTERVAL15s
Agent timeout--agent-timeoutMULTICA_AGENT_TIMEOUT0（不限制，由看门狗兜底）
Codex app-server handshake timeout--codex-handshake-timeoutMULTICA_CODEX_HANDSHAKE_TIMEOUT30s
Max concurrent tasks--max-concurrent-tasksMULTICA_DAEMON_MAX_CONCURRENT_TASKS20
Daemon ID--daemon-idMULTICA_DAEMON_IDhostname
Device name--device-nameMULTICA_DAEMON_DEVICE_NAMEhostname
Runtime name--runtime-nameMULTICA_AGENT_RUNTIME_NAMELocal Agent
Workspaces root—MULTICA_WORKSPACES_ROOT~/multica_workspaces

Agent-specific overrides:

VariableDescription
MULTICA_CLAUDE_PATHCustom path to the claude binary
MULTICA_CLAUDE_MODELOverride the Claude model used
MULTICA_CODEX_PATHCustom path to the codex binary
MULTICA_CODEX_MODELOverride the Codex model used
MULTICA_CODEBUDDY_PATHCustom path to the codebuddy binary
MULTICA_CODEBUDDY_MODELOverride the CodeBuddy model used
MULTICA_ANTIGRAVITY_PATHCustom path to the agy binary
MULTICA_ANTIGRAVITY_MODELOverride the Antigravity model used
MULTICA_OPENCODE_PATHCustom path to the opencode binary
MULTICA_OPENCODE_MODELOverride the OpenCode model used
MULTICA_OPENCLAW_PATHCustom path to the openclaw binary
MULTICA_OPENCLAW_MODELOverride the OpenClaw model used
MULTICA_HERMES_PATHCustom path to the hermes binary
MULTICA_HERMES_MODELOverride the Hermes model used
MULTICA_GEMINI_PATHCustom path to the gemini binary
MULTICA_GEMINI_MODELOverride the Gemini model used
MULTICA_PI_PATHCustom path to the pi binary
MULTICA_PI_MODELOverride the Pi model used
MULTICA_CURSOR_PATHCustom path to the cursor-agent binary
MULTICA_CURSOR_MODELOverride the Cursor model used
MULTICA_KIMI_PATHCustom path to the kimi binary
MULTICA_KIMI_MODELOverride the Kimi model used
MULTICA_KIRO_PATHCustom path to the kiro-cli binary
MULTICA_KIRO_MODELOverride the Kiro model used
MULTICA_QODER_PATHCustom path to the qodercli binary
MULTICA_QODER_MODELOverride the Qoder model used
MULTICA_TRAECLI_PATHCustom path to the traecli binary
MULTICA_TRAECLI_MODELOverride the Trae CLI model used

Self-Hosted Server

When connecting to a self-hosted Multiflow instance, point the CLI to your server before logging in:

export MULTICA_APP_URL=https://app.example.com
export MULTICA_SERVER_URL=wss://api.example.com/ws

multiflow login
multiflow daemon start

Or set them persistently:

multiflow config set app_url https://app.example.com
multiflow config set server_url wss://api.example.com/ws

Profiles

Profiles let you run multiple daemons on the same machine — for example, one for production and one for a staging server.

# Set up a staging profile
multiflow setup self-host --profile staging --server-url https://api-staging.example.com --app-url https://staging.example.com

# Start its daemon
multiflow daemon start --profile staging

# Default profile runs separately
multiflow daemon start

Each profile gets its own config directory (~/.multica/profiles/<name>/), daemon state, health port, and workspace root.

Workspaces

List Workspaces

multiflow workspace list

Watched workspaces are marked with *. The daemon only processes tasks for watched workspaces.

Watch / Unwatch

multiflow workspace watch <workspace-id>
multiflow workspace unwatch <workspace-id>

Get Details

multiflow workspace get <workspace-id>
multiflow workspace get <workspace-id> --output json

List Members

multiflow workspace member list <workspace-id>

Invite Member

按邮箱邀请成员加入 workspace。被邀请人会收到一条待接受（pending）的邀请，接受后才真正加入——不是即时添加。workspace 参数可选，接受完整 UUID、slug 或短 UUID 前缀（≥4 位十六进制），省略时用当前默认 workspace（--workspace-id / MULTICA_WORKSPACE_ID / profile 默认）。

multiflow workspace member invite alice@example.com
multiflow workspace member invite alice@example.com <workspace-id> --role admin
multiflow workspace member invite alice@example.com --output json

--role 默认 member，可传 admin；不允许邀请 owner。目前只支持邮箱标识（不做用户名/工号解析）。若该邮箱已是成员或已有待接受邀请，服务端返回 409 并给出对应提示。

Update Workspace

需要 admin 或 owner 权限。所有字段都是部分更新（PATCH 语义）：未传的字段保持不变。

multiflow workspace update <workspace-id> --name "Acme Eng"
multiflow workspace update <workspace-id> \
  --description "Engineering team workspace" \
  --issue-prefix ENG

长文本走 stdin（保留换行/反斜杠）：

cat <<'CTX' | multiflow workspace update <workspace-id> --context-stdin
我们是一支 5 人 AI-native 团队。
工作语言：中文 + 英文混合。
CTX

可编辑字段：--name、--description / --description-stdin、--context / --context-stdin、--issue-prefix。slug 创建后只读，不暴露在 CLI。--description 与 --description-stdin（以及 context 同名对）互斥。未传任何字段 flag 时命令拒绝执行，避免空 PATCH 触发无意义的 workspace 更新事件。--issue-prefix "" 也会被拒绝：当前后端在 prefix 为空时静默跳过该字段，CLI 在本地拦下避免“看似成功的 no-op”。

Issues

List Issues

multiflow issue list
multiflow issue list --status in_progress
multiflow issue list --priority urgent --assignee "Agent Name"
multiflow issue list --assignee-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
multiflow issue list --full-id
multiflow issue list --limit 20 --output json

表格输出默认显示可直接复制到后续命令的 issue KEY（例如 MUL-123）；需要完整 UUID 时使用 --full-id。Available filters: --status, --priority, --assignee / --assignee-id, --project, --limit. 在重名 workspace 下用 --assignee-id <uuid> 可以精确锁定一个成员或 agent。

Get Issue

multiflow issue get MUL-123
multiflow issue get <uuid>
multiflow issue get <id> --output json

<id> 同时接受 issue key（multiflow issue list 表格里直接显示，例如 MUL-123）和完整 UUID（给 list 加 --full-id 可显示）。同样的规则适用于下面 update / assign / status / comment / subscriber / runs 等接受 <id> 的命令。

Create Issue

multiflow issue create --title "Fix login bug" --description "..." --priority high --assignee "Lambda"
multiflow issue create --title "Fix login bug" --assignee-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d

Flags: --title (required), --description, --status, --priority, --assignee / --assignee-id, --parent, --project, --due-date. 脚本里如果已经拿到了 UUID（例如来自 multiflow workspace member list --output json），传 --assignee-id <uuid>（与 --assignee 互斥）以精确锁定。

Update Issue

multiflow issue update <id> --title "New title" --priority urgent

Assign Issue

multiflow issue assign <id> --to "Lambda"
multiflow issue assign <id> --to-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
multiflow issue assign <id> --unassign

--to-id <uuid>（与 --to 互斥）按 UUID 精确分配；适合重名 workspace 下脚本化场景。

Change Status

multiflow issue status <id> in_progress

Valid statuses: backlog, todo, in_progress, in_review, done, blocked, cancelled.

Comments

# List comments
multiflow issue comment list <issue-id>

# Add a comment
multiflow issue comment add <issue-id> --content "Looks good, merging now"

# Reply to a specific comment
multiflow issue comment add <issue-id> --parent <comment-id> --content "Thanks!"

# Delete a comment
multiflow issue comment delete <comment-id>

Execution History

# List all execution runs for an issue
multiflow issue runs <issue-id>
multiflow issue runs <issue-id> --full-id
multiflow issue runs <issue-id> --output json

# View messages for a specific execution run
multiflow issue run-messages <task-id>
multiflow issue run-messages <short-task-id> --issue <issue-id>
multiflow issue run-messages <task-id> --output json

# Incremental fetch (only messages after a given sequence number)
multiflow issue run-messages <task-id> --since 42 --output json

runs 的表格输出默认显示 task UUID 短前缀；需要完整 task UUID 时使用 --full-id。run-messages 可直接接受完整 task UUID；从 runs 表格复制短前缀时需要同时传 --issue <issue-id>，CLI 只会在该 issue 的 runs 内解析。

Projects

Projects group related issues (e.g. a sprint, an epic, a workstream). Every project belongs to a workspace and can optionally have a lead (member or agent).

List Projects

multiflow project list
multiflow project list --status in_progress
multiflow project list --output json

Available filters: --status.

Get Project

multiflow project get <id>
multiflow project get <id> --output json

Create Project

multiflow project create --title "2026 Week 16 Sprint" --icon "🏃" --lead "Lambda"

Flags: --title (required), --description, --status, --icon, --lead, --start-date, --due-date。日期是日历日（YYYY-MM-DD）。

Update Project

multiflow project update <id> --title "New title" --status in_progress
multiflow project update <id> --lead "Lambda"
multiflow project update <id> --due-date 2026-04-15

Flags: --title, --description, --status, --icon, --lead, --start-date, --due-date。日期传空字符串（如 --start-date ""）可清除。

Change Status

multiflow project status <id> in_progress

Valid statuses: planned, in_progress, paused, completed, cancelled.

Delete Project

multiflow project delete <id>

Associating Issues with Projects

Use the --project flag on issue create / issue update to attach an issue to a project, or on issue list to filter issues by project:

multiflow issue create --title "Login bug" --project <project-id>
multiflow issue update <issue-id> --project <project-id>
multiflow issue list --project <project-id>

Configuration

View Config

multiflow config show

Shows config file path, server URL, app URL, and default workspace.

Set Values

multiflow config set server_url wss://api.example.com/ws
multiflow config set app_url https://app.example.com
multiflow config set workspace_id <workspace-id>

Other Commands

multiflow version              # Show CLI version and commit hash
multiflow update               # Update to latest version
multiflow agent list           # List agents in the current workspace

Output Formats

Most commands support --output with two formats:

  • table — human-readable table (default for list commands)
  • json — structured JSON (useful for scripting and automation)
multiflow issue list --output json
multiflow daemon status --output json

本页目录

AuthenticationBrowser LoginToken LoginCheck StatusLogoutAgent DaemonStartStopStatusLogsSupported AgentsHow It WorksConfigurationSelf-Hosted ServerProfilesWorkspacesList WorkspacesWatch / UnwatchGet DetailsList MembersInvite MemberUpdate WorkspaceIssuesList IssuesGet IssueCreate IssueUpdate IssueAssign IssueChange StatusCommentsExecution HistoryProjectsList ProjectsGet ProjectCreate ProjectUpdate ProjectChange StatusDelete ProjectAssociating Issues with ProjectsConfigurationView ConfigSet ValuesOther CommandsOutput Formats