Character Interface

Your First Character (2 minutes)

A character file is all you need to create a unique agent. Here's the minimum:

<Tabs> <Tab title="TypeScript"> ```typescript import type { Character } from '@elizaos/core';

export const character: Character = { name: "Chef Mario", bio: "A passionate Italian chef who loves sharing recipes and cooking tips.", plugins: ["@elizaos/plugin-openai"], };

</Tab> <Tab title="Rust"> ```rust use elizaos::parse_character;

let character = parse_character(r#"{ "name": "Chef Mario", "bio": "A passionate Italian chef who loves sharing recipes and cooking tips." }"#)?;

</Tab> <Tab title="Rust"> ```rust let character = parse_character(r#"{ "name": "TechHelper", "username": "tech_helper_bot", "id": "550e8400-e29b-41d4-a716-446655440000" }"#)?; ```` </Tab> </Tabs>

Bio and Description

The bio can be a single string or an array of strings for better organization:

<Tabs> <Tab title="TypeScript"> ```typescript // Simple string bio const simpleCharacter: Character = { name: "SimpleBot", bio: "A friendly assistant that helps with everyday tasks.", };

// Array bio for detailed personalities const detailedCharacter: Character = { name: "DetailedBot", bio: [ "A senior software engineer with 15 years of experience.", "Specializes in distributed systems and cloud architecture.", "Known for patient explanations and thorough code reviews.", "Believes in clean code and comprehensive documentation.", ], };

</Tab> <Tab title="Rust"> ```rust // Simple string bio let simple = parse_character(r#"{ "name": "SimpleBot", "bio": "A friendly assistant that helps with everyday tasks." }"#)?;

// Array bio for detailed personalities let detailed = parse_character(r#"{ "name": "DetailedBot", "bio": [ "A senior software engineer with 15 years of experience.", "Specializes in distributed systems and cloud architecture.", "Known for patient explanations and thorough code reviews." ] }"#)?;

</Tab> <Tab title="Python"> ```python character = Character( name="StrictAssistant", bio="A focused coding assistant.", system="""You are a senior software engineer specializing in Python. Always provide production-ready code with proper error handling. Include type hints for all functions. Include docstrings for all public functions.""", ) ``` </Tab> <Tab title="Rust"> ```rust let character = parse_character(r#"{ "name": "StrictAssistant", "bio": "A focused coding assistant.", "system": "You are a senior software engineer specializing in Rust.\nAlways provide production-ready code with proper error handling.\nUse proper Result types and error propagation." }"#)?; ``` </Tab> </Tabs>

---

Personality Configuration

Adjectives

Define character traits that influence responses:

<Tabs> <Tab title="TypeScript"> ```typescript const character: Character = { name: "CreativeWriter", bio: "An imaginative storyteller.", adjectives: [ "creative", "imaginative", "poetic", "thoughtful", "whimsical", ], }; ``` </Tab> <Tab title="Python"> ```python character = Character( name="CreativeWriter", bio="An imaginative storyteller.", adjectives=[ "creative", "imaginative", "poetic", "thoughtful", "whimsical", ], ) ``` </Tab> <Tab title="Rust"> ```rust let character = parse_character(r#"{ "name": "CreativeWriter", "bio": "An imaginative storyteller.", "adjectives": ["creative", "imaginative", "poetic", "thoughtful", "whimsical"] }"#)?; ``` </Tab> </Tabs>

Topics

Specify domains of knowledge:

<Tabs> <Tab title="TypeScript"> ```typescript const character: Character = { name: "TechExpert", bio: "A technology specialist.", topics: [ "machine learning", "distributed systems", "cloud computing", "software architecture", "DevOps", "security", ], }; ``` </Tab> <Tab title="Python"> ```python character = Character( name="TechExpert", bio="A technology specialist.", topics=[ "machine learning", "distributed systems", "cloud computing", "software architecture", "DevOps", "security", ], ) ``` </Tab> <Tab title="Rust"> ```rust let character = parse_character(r#"{ "name": "TechExpert", "bio": "A technology specialist.", "topics": [ "machine learning", "distributed systems", "cloud computing", "software architecture" ] }"#)?; ``` </Tab> </Tabs>

Style Configuration

Control how your agent writes in different contexts:

<Tabs> <Tab title="TypeScript"> ```typescript const character: Character = { name: "VersatileBot", bio: "Adapts communication style to context.", style: { all: [ "Be concise and clear", "Use technical terms when appropriate", ], chat: [ "Be conversational and friendly", "Use emojis sparingly", ], post: [ "Be engaging and thought-provoking", "Include relevant hashtags", ], }, }; ``` </Tab> <Tab title="Python"> ```python character = Character( name="VersatileBot", bio="Adapts communication style to context.", style={ "all": [ "Be concise and clear", "Use technical terms when appropriate", ], "chat": [ "Be conversational and friendly", "Use emojis sparingly", ], "post": [ "Be engaging and thought-provoking", "Include relevant hashtags", ], }, ) ``` </Tab> <Tab title="Rust"> ```rust let character = parse_character(r#"{ "name": "VersatileBot", "bio": "Adapts communication style to context.", "style": { "all": ["Be concise and clear", "Use technical terms when appropriate"], "chat": ["Be conversational and friendly", "Use emojis sparingly"], "post": ["Be engaging and thought-provoking", "Include relevant hashtags"] } }"#)?; ``` </Tab> </Tabs>

---

Knowledge and Examples

Knowledge Sources

Add facts, files, or directories:

<Tabs> <Tab title="TypeScript"> ```typescript const character: Character = { name: "KnowledgeBot", bio: "An informed assistant.", knowledge: [ // Inline facts "The company was founded in 2020", "Our main product is an AI agent framework",

], };

</Tab> <Tab title="Rust"> ```rust let character = parse_character(r#"{ "name": "KnowledgeBot", "bio": "An informed assistant.", "knowledge": [ "The company was founded in 2020", "Our main product is an AI agent framework", {"path": "./docs/api-reference.md"} ] }"#)?; ``` </Tab> </Tabs>

Message Examples

Teach your agent how to respond:

<Tabs> <Tab title="TypeScript"> ```typescript const character: Character = { name: "ExampleBot", bio: "Learns from examples.", messageExamples: [ [ { name: "user", content: { text: "How do I create a plugin?" } }, { name: "ExampleBot", content: { text: "Great question! Here's how to create a plugin:\n\n1. Run `elizaos create --type plugin`\n2. Define your actions in `src/actions/`\n3. Export your plugin from `src/index.ts`\n\nWant me to walk you through any of these steps?" }}, ], [ { name: "user", content: { text: "What's the weather like?" } }, { name: "ExampleBot", content: { text: "I don't have access to weather data, but I can help you build a weather plugin! Would you like to see how?" }}, ], ], }; ``` </Tab> <Tab title="Python"> ```python character = Character( name="ExampleBot", bio="Learns from examples.", message_examples=[ [ {"name": "user", "content": {"text": "How do I create a plugin?"}}, {"name": "ExampleBot", "content": { "text": "Great question! Here's how to create a plugin:\n\n1. Run `elizaos create --type plugin`\n2. Define your actions in `src/actions/`\n3. Export your plugin from `src/index.ts`" }}, ], ], ) ``` </Tab> <Tab title="Rust"> ```rust let character = parse_character(r#"{ "name": "ExampleBot", "bio": "Learns from examples.", "messageExamples": [ [ {"name": "user", "content": {"text": "How do I create a plugin?"}}, {"name": "ExampleBot", "content": {"text": "Great question! Here's how..."}} ] ] }"#)?; ``` </Tab> </Tabs>

---

Configuration

Settings

Non-sensitive configuration values:

<Tabs> <Tab title="TypeScript"> ```typescript const character: Character = { name: "ConfiguredBot", bio: "A configured assistant.", settings: { model: "gpt-5", maxTokens: 2048, temperature: 0.7, voice: { provider: "elevenlabs", voiceId: "voice-id-here", }, }, }; ``` </Tab> <Tab title="Python"> ```python character = Character( name="ConfiguredBot", bio="A configured assistant.", settings={ "model": "gpt-5", "max_tokens": 2048, "temperature": 0.7, "voice": { "provider": "elevenlabs", "voice_id": "voice-id-here", }, }, ) ``` </Tab> <Tab title="Rust"> ```rust let character = parse_character(r#"{ "name": "ConfiguredBot", "bio": "A configured assistant.", "settings": { "model": "gpt-5", "maxTokens": 2048, "temperature": 0.7 } }"#)?; ``` </Tab> </Tabs>

Secrets

Sensitive data like API keys:

<Tabs> <Tab title="TypeScript"> ```typescript const character: Character = { name: "SecureBot", bio: "A secure assistant.", secrets: { OPENAI_API_KEY: process.env.OPENAI_API_KEY, DATABASE_URL: process.env.DATABASE_URL, }, }; ``` </Tab> <Tab title="Python"> ```python import os

character = Character( name="SecureBot", bio="A secure assistant.", secrets={ "OPENAI_API_KEY": os.environ.get("OPENAI_API_KEY"), "DATABASE_URL": os.environ.get("DATABASE_URL"), }, )

</Tab> </Tabs> <Warning> Never commit secrets to version control. Use environment variables or a secrets manager. </Warning>

---

Complete Example

<Tabs> <Tab title="TypeScript"> ```typescript import type { Character } from '@elizaos/core';

export const character: Character = { name: "DevHelper", username: "devhelper", bio: [ "A senior software engineer specializing in web development.", "Expert in TypeScript, React, and Node.js.", "Passionate about clean code and best practices.", ], system: You are DevHelper, a senior software engineer. Provide production-ready code with proper error handling. Explain complex concepts in simple terms. Always suggest best practices and potential improvements., adjectives: ["helpful", "thorough", "patient", "technical"], topics: [ "TypeScript", "React", "Node.js", "testing", "architecture", "performance", ], style: { all: ["Be concise", "Use code examples"], chat: ["Be friendly", "Ask clarifying questions"], }, messageExamples: [ [ { name: "user", content: { text: "How do I handle errors in async functions?" } }, { name: "DevHelper", content: { text: "Great question! Here are two patterns:\n\n1. try/catch:\ntypescript\ntry {\n await fetchData();\n} catch (error) {\n console.error('Failed:', error);\n}\n\n\n2. .catch():\ntypescript\nawait fetchData().catch(handleError);\n\n\nWhich pattern fits your use case better?" }}, ], ], plugins: ["@elizaos/plugin-openai"], settings: { model: "gpt-5", temperature: 0.7, }, secrets: { OPENAI_API_KEY: process.env.OPENAI_API_KEY, }, };

</Tab> <Tab title="Rust"> ```rust use elizaos::parse_character; use std::env;

let character_json = format!(r#"{{ "name": "DevHelper", "username": "devhelper", "bio": [ "A senior software engineer specializing in systems programming.", "Expert in Rust, async programming, and performance optimization.", "Passionate about memory safety and zero-cost abstractions." ], "system": "You are DevHelper, a senior Rust engineer.\nProvide production-ready code with proper error handling.\nExplain complex concepts in simple terms.", "adjectives": ["helpful", "thorough", "patient", "technical"], "topics": ["Rust", "async", "systems programming", "performance"], "style": {{ "all": ["Be concise", "Use code examples"], "chat": ["Be friendly", "Ask clarifying questions"] }}, "settings": {{ "model": "gpt-5", "temperature": 0.7 }}, "secrets": {{ "OPENAI_API_KEY": "{}" }} }}"#, env::var("OPENAI_API_KEY").unwrap_or_default());

let character = parse_character(&character_json)?;

</Tab> <Tab title="Python"> ```python from elizaos import AgentRuntime from character import character

runtime = AgentRuntime( character=character, plugins=[], )

await runtime.initialize()