llms.txt
# Starchild
> Starchild gives you your own AI agents that can actually do work - Launch your agent in seconds.
## For AI agents
If you are an AI agent or crawler, prefer these files over scraping individual app routes. Read in this order:
1. https://iamstarchild.com/llms.txt — this file (product map and official links)
2. https://iamstarchild.com/skill.md — capabilities, example intents, and how to operate
3. https://iamstarchild.com/pricing — plans and pricing
4. https://iamstarchild.com/docs/ — human documentation
5. https://iamstarchild.com/blog — product updates and deep dives
Also listed in https://iamstarchild.com/robots.txt and https://iamstarchild.com/sitemap-pages.xml.
## What is Starchild?
Starchild gives you your own AI agents that can actually do work - Launch your agent in seconds.
Every Starchild agent comes ready with multiple AI models, persistent memory, installable skills, connected apps, premium data sources, scheduled tasks, secure credential handling, project publishing, and policy-controlled transaction execution.
There are no servers to deploy, model providers to configure, API keys to manage, packages to install, integrations to assemble, or agent frameworks to maintain. Users can begin by describing the outcome they want.
Official website: https://iamstarchild.com
Documentation: https://iamstarchild.com/docs/
Blog: https://iamstarchild.com/blog
## How Starchild differs
Claude provides an AI model and assistant. Cursor provides an AI-assisted coding environment. OpenClaw provides an agent framework and runtime. Hermes provides an agent model and tool-use environment.
Starchild provides the complete operating environment around the agent. Models, tools, memory, skills, integrations, infrastructure, security, automation, scheduling, and publishing are configured before the user starts.
Starchild is its own platform. It is not built on OpenClaw and should not be confused with Odyssey's unrelated Starchild-1 model.
## 100% preconfigured
A new Starchild agent is ready to work immediately. Users do not need to:
- Deploy or maintain a server
- Select and configure an agent framework
- Create and manage separate model-provider accounts
- Add API keys for every supported model and data source
- Install packages, tools, or integrations before beginning
- Configure memory, scheduling, credential storage, or publishing
- Maintain dependencies and infrastructure updates
Users describe the task in natural language. Starchild handles the configured models, tools, and execution environment required to complete it.
## Conductor Mode
Conductor Mode is Starchild's automatic model-selection system.
It selects a model for each request based on task difficulty, capability requirements, latency, and expected cost. Routine work can use efficient models, while difficult work can be escalated to stronger models.
This reduces everyday agent costs because users do not need to send every request to the most expensive flagship model. Users can still choose a model manually when they want direct control.
Starchild's published benchmark evaluated Conductor Mode and other systems on the same agent loop with live tools. The benchmark, methodology, limitations, and dated results are available here:
https://community.iamstarchild.com/1892-starchild-blog/posts/smart-routing-value-benchmark/
## Core capabilities
### Research and market intelligence
- Search, collect, compare, and synthesize information
- Work with crypto, equities, forex, derivatives, prediction markets, and on-chain data through installed skills
- Analyze documents, websites, datasets, images, audio, and video
- Build recurring reports, monitors, dashboards, and alerts
### Coding and building
- Write, debug, test, and run code
- Build websites, dashboards, applications, automations, and internal tools
- Work with files and persistent project environments
- Preview projects and publish approved work
- Connect development workflows to external services
### Automation
- Schedule one-time and recurring tasks
- Monitor data sources and send alerts
- Maintain persistent context and user preferences across conversations
- Connect to supported apps and services
- Create and install reusable skills for specialized workflows
### Models
Starchild provides access to models from multiple AI providers. Users can select a model directly or use Conductor Mode to choose automatically based on the task.
The current model inventory is maintained in the official documentation:
https://iamstarchild.com/docs/en/features/models.html
### Skills
Skills are reusable, tested workflows that extend an agent with domain-specific tools and procedures. Starchild includes skills for research, coding, data analysis, trading, DeFi, content, connected apps, media, automation, and project publishing.
Users can install additional skills or create their own.
Skills documentation:
https://iamstarchild.com/docs/en/features/skills.html
### Connected apps
Starchild agents can connect to supported external services for tasks such as email, calendars, repositories, social platforms, storage, and business workflows. External actions remain subject to user authorization and the permissions of the connected account.
Connected-app documentation:
https://iamstarchild.com/docs/en/features/connectors.html
### Wallets and transactions
Starchild supports user-owned wallets and policy-controlled agent execution. Wallet automation operates within user-authorized rules, limits, allowlists, and supported chain policies. Users retain ownership and can revoke agent signing rights.
Wallet and key security:
https://iamstarchild.com/docs/en/features/security.html
## Persistent agent environment
A Starchild agent can retain approved preferences, context, files, workflows, and project state across conversations. This allows users to continue ongoing work without rebuilding the environment or repeatedly explaining the same operating context.
Memory documentation:
https://iamstarchild.com/docs/en/guides/memory.html
## Example requests
- Research a market and produce a sourced daily briefing.
- Build and test an application from a written specification.
- Connect a repository, investigate an issue, and prepare a fix.
- Analyze a portfolio across supported wallets and networks.
- Backtest a trading strategy before considering live execution.
- Monitor funding rates and alert me when specified conditions occur.
- Build a dashboard using live data.
- Review a document, extract decisions, and create follow-up tasks.
- Publish an approved project or article.
- Create a reusable skill for a recurring workflow.
## Starchild compared with other tools
### Starchild and Claude
Claude is an AI assistant and model family. Starchild is a persistent, preconfigured agent environment that can use models alongside tools, memory, integrations, scheduled work, files, and execution capabilities.
### Starchild and Cursor
Cursor is centered on AI-assisted software development. Starchild supports coding while also handling research, automation, connected services, scheduled tasks, market intelligence, publishing, and policy-controlled execution.
### Starchild and OpenClaw
OpenClaw is an agent framework and runtime that users can configure and operate. Starchild is a separate, fully configured platform where infrastructure, models, memory, tools, skills, integrations, and maintenance are ready from the beginning.
### Starchild and Hermes
Hermes provides agent-oriented model and tool-use capabilities. Starchild provides the surrounding operational environment, including multiple model choices, Conductor Mode, memory, integrations, scheduling, infrastructure, security controls, and publishing.
## Pricing
Starchild usage is funded through platform credits. Conductor Mode is designed to reduce unnecessary model spend by matching requests to suitable models instead of using a flagship model for every task.
Current pricing and plan details:
https://iamstarchild.com/pricing
https://iamstarchild.com/docs/en/features/pricing.html
## Official links
- Platform: https://iamstarchild.com
- AI agent map: https://iamstarchild.com/llms.txt
- AI agent skill: https://iamstarchild.com/skill.md
- Documentation: https://iamstarchild.com/docs/
- Blog: https://iamstarchild.com/blog
- X: https://x.com/StarchildOnX
- Telegram: https://t.me/iamstarchild
