Does Soloco run its own hosted agents?

No. Soloco coordinates tasks and sends scoped work to the user's own local agents.

Can it work with Claude Code or Codex?

Yes. The core design is CLI dispatch, so local tools can be added behind a common task interface.

How are long contexts handled?

Tasks are split into nodes, then summarized into checkpoints and SOP memory as work is accepted.

Where does local data live?

The client owns local indexing and execution. Cloud sync is used for coordination, identity, billing, and selected metadata.

Need help evaluating a workflow?

Contact support at support@soloco.cloud.

Run multiple companieson your own
Claude / Codex subscription

Tulong Auto · Global StrategyACJMS

New goal

Org · Tulong Auto

Give the AI a goal

Plan itConvene boardStart goal loop

Workbench0 running · 30 agents online

Goals done this month

↑ 1 in progress

Spend this month

$0.00

2.5M tok total

Active agents

● Online

Current goal

No goal running

Running · cycle 62 / 200GOAL-7C3AF2D9

Make Tulong Auto the top auto aftermarket service brand in North America

Progress0%

Recent activity

Two ways, onto your machine

Runs on your machine, on your existing Claude / Codex subscription. Data stays local

Comfortable with CLIOne command to install

$npm install -g soloco

macOS · Windows · Linux

Rather skip the CLICopy the prompt, let AI install it

Install and launch Soloco on my machine — a platform that runs multiple agents on my local Claude / Codex subscription:

1. Make sure Node.js ≥ 20; if not, install it via nvm
2. Install globally: npm install -g soloco
3. Confirm either Claude Code or Codex CLI is installed and logged in (Soloco uses it, no API key needed)
4. Run soloco to launch; if port 3100 is taken, handle it
5. After launch, have me open http://localhost:3100 and briefly explain what's next

Paste into Claude Code, Cursor, or any AI assistant

Auto-connect the AI you already installed

Claude Code and Codex run on your existing subscription — no API key. DeepSeek, Kimi, OpenRouter and other OpenAI-compatible endpoints also work.

Configure AI backendStep 1 of 2

Configure several; pick one as default. Change anytime under Settings → AI backend.

Claude Code (CLI)claude code terminal detected; runs on subscription — recommended2.1.154 (Claude Code)

Codex (CLI)codex terminal detectedcodex-cli 0.114.0

OpenAI-compatible APIDeepSeek / Kimi / OpenRouter / Ollama, etc.

Base URLhttps://api.deepseek.com/v1

API Keysk-...

Default model (optional)deepseek-chat / gpt-4.1-mini

Test & save

Maybe laterNext →

Generate a multi-node team for your goal

Company / orchestration core / divisions / specialists — expands automatically

Org canvasACJMS

Make Tulong Auto the top auto aftermarket service brand in North America31%Running · cycle 62/200Timeline

Tulong Auto · Global

16 agents · cycle 62/200

Orchestration Coreopus-4-8

cycle 47 · spawn 210

UUS Regulatory Researcher

A91F · NA Division

MTD US$12.40

WebSearch·9Read·4+2

HHK/Macau Researcher

6F2A · HK/Macau

MTD US$3.10

WebSearch·5Read·3+1

OOther Markets Researcher

4B7C · International

MTD US$6.80

WebFetch·8Read·5+2

SSEA Researcher

7C3A · SEA

MTD US$7.20

WebFetch·12Read·6+1

CCapital Allocation Lead

5D02 · Finance

MTD US$9.10

Bash·6Write·4+1

TTooling Infra Engineer

9E4D · Tooling

MTD US$5.60

Bash·7Edit·1+3

WWorkflow Builder

W1 · Market Entry

MTD US$15.60

Bash·9Write·8+4

PProposal Reviewer

W1 · Red Team

MTD US$2.69

Read·7Grep·4+1

WWorkflow Builder

W2 · Store SOP

MTD US$13.30

Bash·8Write·6+3

PProposal Reviewer

W2 · Review

MTD US$2.10

Read·6Grep·3+1

WWorkflow Builder

W3 · Supply Chain

MTD US$11.60

Bash·7Write·5+2

WWorkflow Builder

W4 · Digital Platform

MTD US$22.21

Bash·11Edit·4+5

WWorkflow Builder

W5 · Org & Talent

MTD US$11.80

Bash·6Write·5+2

PProposal Reviewer

W5 · Review

MTD US$1.90

Read·5+1

WWorkflow Builder

W10 · KPI Dashboard

MTD US$6.84

Bash·9Write·7+3

This is the “Org canvas” page you see inside Soloco — same sidebar, goal bar, nodes and edges. The nodes are generated for your goal.

The team keeps working
until the artifacts land in your hands

It won't stop until the goal closes. Artifacts land in a folder you can open directly

Goal progressACJMS

Make Tulong Auto the top auto aftermarket service brand in North America: use the US as the anchor, systematically replicate the Tulong workshop-store model to Canada / Mexico, and build a sustainably profitable regional franchise network. Each market × each workflow closes out one by one; artifacts land in workspace, ready to open.

RunningCollaborative$161.05 / $5000.00Cycle 62 · 31%2026/5/30 14:34:21

⏸ Pause⊗ Stop

›_Subprocess Runs47 runs · $161.05

›#47RunningTooling Infra Engineerstrongest$5.6031.7k / 182.0k14:34:21

Migrated single-store financial model into workspace, logging each script run

›#46SuccessCapital Allocation Leadstrongest$9.1062.9k / 330.6k13:58:02

Closed out NA entry financial model v3, ran regional P&L

›#45SuccessWorkflow Builderstrongest$15.6031.9k / 194.8k12:11:40

Finalized W1 market-entry playbook + milestone table

›#44FailedWorkflow Builderstrongest$22.2185.6k / 292.0k11:03:18

W4 digital platform single-shot close-out failed, missing _verification.md, rolled back and re-queued

›#43SuccessUS Regulatory Researcherstrongest$12.4029.0k / 144.8k09:47:55

US auto aftermarket entry regulations review passed

Deliverables92 files

deliverables/US/W1_market_entry/US-W1_market_entry_plan.xlsx22.1 KBDownload

deliverables/US/W2_workshop_sop/gen_us_w2_workshop_sop.py13.3 KBDownload

deliverables/US/W3_procurement/US-W3_procurement_cost.xlsx7.1 KBDownload

deliverables/US/W4_digital_platform/gen_us_w4_digital_platform.py15.5 KBDownload

deliverables/US/W10_kpi_dashboard/US-W10_kpi_dashboard.xlsx10.9 KBDownload

deliverables/US/W1_market_entry/README_US-W1.md3.2 KBDownload