← All docs

🚒 Fleet Architecture

Multi-tenant agent spawning with configurable secrets, customer onboarding, and fleet management.

The Vision

Neo isn't just one agent β€” it's a fleet architecture:

Internal Agents

Full access (Monica, Ray, Neo-Prime)

Customer Agents

Spawned per customer with their own keys

Vertical Agents

Pre-configured for specific domains (Anasto, LotPay)

Architecture

flowchart TB
  FM["πŸ›°οΈ Neo Fleet Manager
spawn Β· configure Β· monitor"] subgraph TYPES["Agent Fleet"] direction LR P["πŸ” Neo-Prime (Internal)
full secrets Β· full memory Β· all skills"] C["πŸ‘€ Neo-LW (Customer)
customer secrets · automotive skills"] V["🏒 Neo-Anasto (Vertical)
biomedical skills · customer secrets"] end subgraph INFRA["🧱 Shared Infrastructure"] direction LR R["Railway
hosting"] O["OpenClaw
runtime"] S["soul.py
memory"] SB["Supabase
fleet config"] CH["Telegram / Discord / API"] end FM --> P FM --> C FM --> V P --> INFRA C --> INFRA V --> INFRA

Agent Types

πŸ”

Internal Agents (Full Trust)

Examples: Monica, Ray, Neo-Prime

  • β€’ Secrets: All our API keys, tokens, credentials
  • β€’ Memory: Full access to MEMORY.md, all projects
  • β€’ Skills: Everything
πŸ‘€

Customer Agents (Isolated)

Examples: Neo-LW (LW Automotive), Neo-CustomerX

  • β€’ Secrets: Customer provides their own (GitHub PAT, OpenAI key, etc.)
  • β€’ Memory: Fresh start OR seeded with domain knowledge (no personal data)
  • β€’ Skills: Configured per customer needs
🏒

Vertical Agents (Domain-Specific)

Examples: Neo-Anasto (biomedical), Neo-Finance, Neo-Legal

  • β€’ Secrets: Mix (our infra keys + customer API keys)
  • β€’ Memory: Seeded with domain expertise
  • β€’ Skills: Pre-loaded domain skills

Config Layers

Different agents need different credentials. Solution: layered config inheritance.

# Base config (all agents inherit)
base:
  runtime: openclaw
  hosting: railway
  memory: soul.py
  
# Internal overlay (add our secrets)
internal:
  inherit: base
  secrets:
    GITHUB_TOKEN: "gho_OUR_TOKEN"
    OPENAI_API_KEY: "sk-OUR_KEY"
  memory:
    bootstrap: "full"  # Load all MEMORY.md
    
# Customer overlay (their secrets)
customer:
  inherit: base
  secrets:
    GITHUB_TOKEN: "${CUSTOMER_GITHUB_TOKEN}"
    OPENAI_API_KEY: "${CUSTOMER_OPENAI_KEY}"
  memory:
    bootstrap: "domain"  # Domain knowledge only
    
# Vertical overlay (domain + customer)
vertical-anasto:
  inherit: customer
  skills:
    - biomedical-research
    - clinical-trials
    - medical-devices
  memory:
    seed: "anasto-domain-knowledge"

Customer Onboarding

Set Up Your AI Developer Agent

1. GitHub Access

Create a GitHub Personal Access Token with scopes:

β˜‘ repo β˜‘ read:user β˜‘ user:email

2. AI Model Access

β—‹ Use Copilot (via GitHub)
β—‹ Use OpenAI directly
β—‹ Use Anthropic directly

3. Communication Channel

β—‹ Telegram (recommended)
β—‹ Discord
Create My Agent

Spawn Flow

πŸ“
Customer fills wizard
β†’
βœ…
Validate credentials
β†’
πŸš€
Create Railway service
β†’
πŸ”‘
Inject secrets
β†’
πŸ’¬
Agent ready!

Fleet Dashboard

Neo Fleet Manager

Admin Panel

Active Agents (4)

AgentCustomerStatusLast Active
Neo-PrimeInternal🟒 Up2 min ago
Neo-LWLW Auto🟒 Up15 min ago
Neo-Anasto1CJ Shores🟒 Up1 hr ago
Neo-TrialDemo🟑 Idle3 days ago
+ Spawn New Agent Manage Templates View Logs

Railway Compute

$47.23 / mo

API Calls

12,847

OpenAI 8.2K Β· Copilot 4.6K

Storage

2.3 GB

Security

πŸ” Secrets

  • β€’ Never log secrets
  • β€’ Hash for verification only
  • β€’ Encrypt at rest
  • β€’ Rotate regularly

πŸ”’ Isolation

  • β€’ Separate Railway services
  • β€’ Separate memory per agent
  • β€’ No cross-customer access
  • β€’ Audit logging

πŸ‘€ Customer Data

  • β€’ Customer owns their data
  • β€’ Export/delete anytime
  • β€’ We don't access repos
  • β€’ GDPR-ready

Pricing Model

Plan Agents Features Price
Trial1Basic skills, 7-day limitFree
Pro3Full skills, priority support$49/mo
Team10+ Fleet dashboard, SSO$199/mo
EnterpriseUnlimited+ Custom skills, SLACustom
← Back to Neo