MOBILE CONTROL PLANE FOR AI CODING AGENTS

Plan, run, and ship with your AI agents from anywhere.

Self-hosted on your VM / cloud.
Use your existing AI subscriptions.
Launch agents from your phone.
Review and ship from the lock screen.

See it in action

[claude] reading workers/pool.js

[codex] writing routes/auth.ts

[claude] bash: npm test

[kimi] reading lib/jwt.ts

[claude] writing middleware.ts

[codex] bash: git commit -m feat

[claude] PR opened: #142

[gemini] reviewing changes

[claude] bash: lint --fix

[codex] writing test fixtures

[claude] reading workers/pool.js

[codex] writing routes/auth.ts

[claude] bash: npm test

[kimi] reading lib/jwt.ts

[claude] writing middleware.ts

[codex] bash: git commit -m feat

[claude] PR opened: #142

[gemini] reviewing changes

[claude] bash: lint --fix

[codex] writing test fixtures

2:11

5G ● 69%

Syndic

All 4

1 Active

1 Idle

1 Input

luna

luna/add-pagination · just now

Add pagination to user API

● RUNNING

Analysing schema and pagination strategy · 1m

Refactor auth middleware

AWAITING_REVIEW

PR #4 awaiting review · 22m

Tests passing · 3 commits

8 files changed +261 −58

Plan the job queue backend

● NEEDS_INPUT

Agent asked a question · now

"Should I use Redis or Postgres for the job queue?"

apex

apex/redis-fix · no recent activity

Fix Redis connection pool

DONE

Connection pool patched and tests passing · 41m

Fleet

2Inbox

Review

Terminal

Settings

Add pagination to user API

↳ luna · add-pagination · claude

● LIVE

Add pagination to user API

running · 0:12

The users endpoint accepts pagination params but never returns a cursor.

☑Add stable cursor (created_at + id)

☑Validate limit param

☑Cover empty-page case

☐Open PR for review

read_fileroutes/users.ts

bashpnpm test

✎edit · routes/users.ts

const limit = clamp(query.limit, 1, 100);

const cursor = parseCursor(query.cursor);

const users = await listUsers({ limit, cursor });

return json(users);

return json({ data: users, nextCursor });

});

tests passing · 28 tool calls

API stays backwards compatible.

Can you also open a PR for this

running · 0:04

awaiting review

fix: add cursor pagination

✎+261−583 files

Review diff

Approve & merge

Describe the task… or type / for commands

opus 4.7

■

↑

Fleet

2Inbox

Review

Terminal

Settings

The

I'm

⇧

⌫

123

☺

⏎

⌘

⌕

← PR #4 · 3 files

routes/users.ts

+261 −58

File 1 of 3 · 2 hunks

‹

›

☰

const limit = clamp(query.limit, 1, 100);

+ const cursor = parseCursor(query.cursor);

+ const users = await listUsers({ limit, cursor });

- return json(users);

+ return json({ data: users, nextCursor });

});

Approve

Comment

Reject

Fleet

2Inbox

Review

Terminal

Settings

01 — WORKFLOW

Dispatch · Monitor · Review · Audit Log.

Dispatch from anywhere.

Describe a task. Pick a repo. Pick a CLI. Tap launch. Syndic spins up a worktree on your server in under 30 seconds.

Voice or text input
Templates for repeat workflows
Up to 50 concurrent agents
Notifications when input is needed

Monitor every move.

A live activity feed shows every tool call, every file read, every command. Tap any agent to see what it's doing right now.

Live activity feed for every tool call
Quick glance widget on lock screen
Push notifications when input is needed
Per-agent timer and turn count

Review and merge.

Inspect the diff, leave comments, request changes, or merge directly. From your phone. From the lock screen. From anywhere.

Diff viewer with syntax highlighting
Inline comments and change requests
One-tap merge
CI status in line

2:11

5G ● 69%

‹ Refactor auth mid… Changed Files PR #4 ›

3 COMMITS AGAINST MAIN

a1b2c3d feat: add cursor pagination to user endpoint 24m

e4f5g6h test: add pagination integration tests 23m

i7j8k9l refactor: extract pagination helpers 24m

8 FILES CHANGED

📁 src/users 3

users.controller.ts +42 -8 ›

users.service.ts +38 -4 ›

users.module.ts +3 -2 ›

📁 src/dto 2

pagination.dto.ts +28 ›

cursor.dto.ts +22 ›

📁 test 2

users.e2e.ts +88 ›

pagination.e2e.ts +40 ›

Fleet

2 Inbox

Review

Terminal

Settings

See everything. Audit anything.

A complete, searchable history of every action your fleet takes. Every message, every workspace, every config change, attributed and timestamped.

Full operation log across agents, workspaces, repos, and system
Search and filter by source or actor
Today, yesterday, and earlier views in one timeline
Tap any entry for the underlying details

2:11

5G ● 69%

‹ Audit Log

Search operations…

All

Agent

Workspace

Repo

System

TODAY

Sent message 19:35

Plan the job queue backend

Sent reply to waiting agent

Triggered security scan 19:31

Demo security scan completed

Updated repository settings 19:26

Updated agent instructions path

Created workspace 19:20

Add pagination to user API

Created workspace on luna/add-pagination

Stopped agent 19:14

Polish onboarding empty states

Stop ignored, workspace was complete

Stored GitHub token 18:56

Stored GitHub token for Maya Chen

YESTERDAY

Added repository 17:37

Added demo repository apex

Fleet

2 Inbox

Review

Terminal

Settings

Your agents, your infra, connected.

Syndic connects the agent CLI you already use to the VM, cloud box, or server where the work should run.

First-party model agents

Support the coding CLIs backed by their own model families.

Claude Code

Codex

Gemini CLI

Kimi CLI

Runs on your infra

Self-host on your VM, cloud box, or server.

AWS

GCP

Hetzner

DigitalOcean

Fly

Any Linux VM

02 — WHY SYNDIC

What cloud agents are missing.

Most tools run agents. Syndic adds the missing control layer: mobile access, your subscriptions, your compute, and self-hosting.

Typical cloud agents

Their tokensYour data, their control.

Their infraTheir cloud, their rules.

Their workflowTheir pipeline, their process.

Their limitsTheir quotas, their ceilings.

With Syndic

Your subscriptionsUse what you already pay for.

Your VMRuns on your infra, your way.

Your repoYour code, your branches.

Your review flowYou plan, dispatch, review, merge.

Your phonePlan, dispatch, review on the go.

│ ● a3f9d2e  feat: add JWT refresh token rotation
│ │
│ ○ b81c4f1  refactor: extract auth middleware
├─╯
● │ c5e2a30  fix: connection pool timeout on retry
│ │
○ │ d94b7e8  feat: goal mode lifecycle renderer
│ ╭─┤
│ ● │ e120f4c  chore: update deps, bump minor
│ │ │
│ ○ │ f3d8a17  feat: worker fleet status endpoint
│   │
●   │ 1a4c9b6  fix: race condition in dispatch queue
│   │
○   ╰─● 2b7f3e4  feat: sandboxed exec container init
│     │
●     ○ 3c8d2a9  refactor: split repo into packages
│     │
│   ╭─╯
○   ● 4e1b5f8  fix: mobile keyboard overlap on iOS
│   │
│   ○ 5f2c6d7  feat: diff viewer syntax highlight
├───╯
● │ 6a3d7e2  perf: streaming response buffering
│ │
○ │ 7b4e8f1  test: add integration tests for goals
│ ╭╯
● ● 8c5f9a0  feat: multi-agent parallel dispatch
│ │
○ │ 9d6a0b3  docs: update self-hosting guide
│ │
●─╯ ae7b1c4  release: v0.4.0

2:11

5G ● 69%

‹

New Goal

⋯

luna · goal intake · world-class goal ux

Plan first

Run now

Repository

luna

›

Goal

Stabilize auth rollout before the mobile release

Rough intent first. Plan your goals before you run them.

Context · 2 items

auth-rollout-notes.mdDOC

src/auth/middleware.tsFILE

Model

✦Claude 4 · planning layer

›

Plan goal

‹

Goal planning

⋯

luna · model-assisted planning · before execution

Planning pass

using the right models to research, scope, and de-risk the goal

› research the current auth flow and rollout notes

› scope the safe phases before any worker runs

› flag migration risks and required checks

planning layer only — execution starts after approval

Building launchable goal

research · scope · risks · checks

‹

Goal editor

⋯

v1 · launchable goal · 3 phases · 4 tasks

Auth rollout stabilization

notes-style goal with phases, tasks, and checks

01Phase 1 · audit cookie paths and JWT issuance

02Phase 2 · ship middleware swap behind checks

03Task · update the mobile + web auth handoff

04Checks · regression suite, rollback note, release review

Phases ready 4 tasks 3 checks

Launch goal

‹

Launch goal

⊟

orchestrator ready · 3 workers · parallel

Dispatch

One by one Parallel · 3 Manual review

Orchestrator

READY

sequencing phases · reviews · unblock routing

Research + scope

CLAUDE

new worktree · auth-rollout/research-scope

Middleware + checks

CODEX

new worktree · auth-rollout/middleware-checks

Client handoff

KIMI

new worktree · auth-rollout/client-handoff

Launch goal

‹

Auth rollout stabilization

⋯

luna · stream live · phase 1 running

Goal

Stream

Files

Pause Unblock Review

Phase 1streaming · parallel

Orchestrator status

LIVE

claude · syncing worker progress and reviews

•09:41 scoped risks and checks

•09:43 launched middleware and client worktrees

Middleware + checks

RUN

codex · validating auth guards and rollback checks

Client handoff

RUN

kimi · updating token refresh handoff

Research + scope

REVIEW

claude · ready for goal review before phase 2

Phase 2queued after review

Release rollout

QUEUE

waiting for unblock + approval

‹

Goal fleet

⋯

3 goals · 1 complete · 1 review · 1 running

Fleet

Stream

Files

Auth rollout stabilization

COMPLETE

claude · codex · kimi

Mobile session cleanup

NEEDS REVIEW

codex · waiting on release checklist

API access audit

RUNNING

kimi · collecting final evidence for review

03 — GOAL MODE

Set the goal.
Shape the path.

Plan with the right models before you launch long-running agent work.

Turn rough intent into runnable goals
Use the right models before execution starts
Shape phases, tasks, checks, and blockers
Launch, pause, unblock, and review from your phone

~ laptop · syndic

~ laptop $ syndic skill push ./pr-review-bot --target both
  ↗ uploading bundle ··········· 142 KB
  ✓ pr-review-bot → prod-runner-01 · synced

~ laptop $ syndic configure-agents --all
  ✓ claude + codex configs → prod-runner-01 · synced

~ laptop $ syndic ws spawn luna --prompt "add rate limiting" --harness claude
  ✓ workspace ws-a3f82b created on prod-runner-01
  ↗ claude · opus · PR #128 ready

~ laptop $ █

04 — CONTROL

Command your entire fleet from desktop.

Push your skills and agent configs to any remote machine in one command, then spawn, watch, and stop tasks on it without leaving your terminal.

Push skills and agent configs from local to remote in one command
Same setup on every machine you touch
Spawn and stop tasks on any server from one place
Stream live logs and diffs back to your desktop

05 — COMPUTE

Put the heavy work in the cloud.

Run agents on rented VMs from your provider of choice. Cheap when idle, powerful when needed, disposable when done.

Rent any VM

Use Hetzner, Fly, AWS, GCP, Azure, or your own box. Syndic runs where your compute already lives.

Scale per task

Start small for everyday work, then resize or swap machines when a plan needs more cores, memory, or parallelism.

Run in parallel

Spin up multiple agents without cooking your laptop. Keep local work light while cloud machines do the heavy lifting.

06 — PRICING

Ship more. From anywhere.

Run the server free. Pay only for the mobile experience.

Free

$0 / forever

Everything you need
to start dispatching.

2 repositories
4 concurrent agents
Dispatch, monitor, review
Diff viewer + merge

Get Started

Pro

$59 / year

Everything in Free, plus full
control over your fleet.

Unlimited repos + agents
Per-agent configuration
Activity monitor
Audit log
Automated security scans
Priority support
Terminal access to your VM

Unlock Pro

┌ syndic server ───────────────────────────────────────┐
│ status: ready        tunnel: secure                  │
│ pairing: active      agents: online                  │
│ routes: /dispatch    /review    /merge               │
└──────────────────────────────────────────────────────┘

Get started in one command.

$ curl -fsSL https://get.syndic.dev | sh

Star on GitHub Download on the App Store

Plan, run, and ship with your AI agents from anywhere.

Dispatch · Monitor · Review · Audit Log.

Dispatch from anywhere.

Monitor every move.

Review and merge.

See everything. Audit anything.

Your agents, your infra, connected.

What cloud agents are missing.

Typical cloud agents

With Syndic

Set the goal.Shape the path.

Command your entire fleet from desktop.

Put the heavy work in the cloud.

Rent any VM

Scale per task

Run in parallel

Ship more. From anywhere.

Get started in one command.

Set the goal.
Shape the path.