Why Ethos?

Ethos makes different tradeoffs than other agent frameworks. Here's an honest comparison.

Feature comparison

Feature	Ethos	LangChain	CrewAI	AutoGen
Personality as structure (not a prompt string)	✅	❌	❌	❌
Swap LLM provider without code changes	✅	~	~	~
TypeScript-first interface contracts	✅	❌	❌	❌
Multi-platform shared sessions	✅	❌	❌	❌
Memory scope per personality	✅	❌	❌	❌
Tool access per personality	✅	~	~	❌
Zero-dependency interface package	✅	❌	❌	❌
Session persistence across restarts	✅	~	~	~

✅ = full support · ~ = partial · ❌ = not supported

The key differences

1. Personality is a structural component

In LangChain, CrewAI, and AutoGen, "personality" means setting a system prompt string. Changing it changes how the model responds — but nothing else in the system changes.

In Ethos, a personality is a directory of files. Swapping it changes:

• The system prompt (via ETHOS.md)
• The tool access (via toolset.yaml)
• The memory scope (via memoryScope in config.yaml)
• The model being used (via model in config.yaml)

All four change atomically. You can't accidentally run the engineer personality's tools with the reviewer's restricted toolset.

2. TypeScript-first, interface-driven

Every extension point in Ethos is a typed interface in @ethosagent/types:

interface LLMProvider {
  complete(messages: Message[], options: CompletionOptions): AsyncIterable<CompletionChunk>
}

interface SessionStore {
  getMessages(sessionId: string, options?: { limit?: number }): Promise<Message[]>
  addMessage(sessionId: string, message: Message): Promise<void>
}

These interfaces have zero dependencies. Any package can implement them. Core never imports concrete implementations.

Python frameworks pass dicts and strings. TypeScript catches mistakes at compile time, not at runtime when the agent is mid-task.

3. Swap everything

Component	Interface	Default	Alternatives
LLM	LLMProvider	Anthropic, OpenAI-compat	Any HTTP-based LLM
Sessions	SessionStore	SQLite (WAL+FTS5)	Redis, Postgres, in-memory
Memory	MemoryProvider	Markdown files	Any storage
Platform	PlatformAdapter	CLI	Telegram, Discord, Slack
Personalities	PersonalityRegistry	File system	Remote registry

LangChain has swap-ability in theory; in practice, changing the underlying LLM requires touching provider-specific abstractions throughout. In Ethos, LLMProvider is one interface with one method.

4. Multi-platform, shared sessions

The same agent — same personality, same memory, same session history — runs across CLI, Telegram, Discord, and Slack. A user can start a conversation on Telegram and continue it on CLI. The session key determines continuity, not the platform.

Other frameworks are designed for single-platform deployment. Adding a second platform typically requires duplicating configuration and state management.

When Ethos isn't the right choice

Use LangChain if: You're building complex multi-step pipelines with many chained operations, or you need the large ecosystem of pre-built integrations.

Use CrewAI if: You're building multi-agent systems where several agents collaborate on a task.

Use AutoGen if: You need sophisticated multi-agent conversation patterns or code execution sandboxes.

Use Ethos if: You're building an interactive agent that a real user talks to, you care about TypeScript correctness, you want to run on multiple platforms, or personality and memory isolation matter to your use case.