IDENTITY.md defines your OpenClaw agent's public persona — its name, emoji, creature type, avatar description, and display identity. This file controls how your agent introduces itself and how it appears in multi-agent setups.
OpenClaw IDENTITY.md Guide: Name and Brand Your AI Agent
Trusted by 194+ business owners who configured OpenClaw in our live workshop
What Is IDENTITY.md?
IDENTITY.md is the OpenClaw configuration file that controls your agent's external presentation. It answers the question: "How does this agent look and introduce itself to the world?" Every field in IDENTITY.md is about perception — the name users see, the emoji that identifies the agent in chat, the creature type that classifies its role, and the avatar description that gives it a visual presence.
The file lives at ~/.openclaw/workspace/IDENTITY.md alongside the rest of OpenClaw's identity architecture. It is a plain Markdown file with simple key-value fields — no JSON, no YAML, no proprietary format. You can open it in VS Code, Vim, Notepad, or any text editor. Because it is a standard file, you can version-control it with Git, share it across teams, and roll back changes at any time.
IDENTITY.md is distinct from SOUL.md (which defines the agent's internal philosophy, values, and reasoning style) and STYLE.md (which controls formatting and writing tone). Think of it this way: SOUL.md is who the agent is, STYLE.md is how the agent writes, and IDENTITY.md is how the agent presents itself. All three work together, but each has a focused, non-overlapping role.
The six core fields in IDENTITY.md are name, emoji, creatureType, pronouns, avatar, and tagline. Beyond these fields, the file includes two optional sections — Introduction Style and Multi-Agent Display — that let you fine-tune how the agent greets users and how it labels its messages in group chats with multiple agents.
Name & Emoji
Set a memorable display name and a single emoji for instant visual identification across every interface.
Creature Type
Classify your agent as an assistant, advisor, companion, specialist, or sentinel to define its role.
Avatar & Tagline
Provide a visual description for image generation and a one-line pitch for agent directories.
Full IDENTITY.md Template
Copy this template, save it to ~/.openclaw/workspace/IDENTITY.md, and customize each field. Every line is optional — start with the fields that matter most to you and add more over time.
# IDENTITY
name: "Atlas"
emoji: "🏛️"
creatureType: "advisor"
pronouns: "they/them"
avatar: "A marble bust of a classical philosopher with glowing blue circuit patterns"
tagline: "Your strategic thinking partner"
## Introduction Style
When greeting users, introduce yourself by name and emoji.
Keep introductions brief — one sentence max.
## Multi-Agent Display
In group chats, always prefix messages with your emoji and name.Fields
6
name, emoji, creatureType, pronouns, avatar, tagline
Format
Plain Markdown
Key-value pairs and section headings
Location
~/.openclaw/workspace/
Same directory as SOUL.md and AGENTS.md
Understanding Each IDENTITY.md Field
Each field in IDENTITY.md serves a specific purpose. Here is what each one does, why it matters, and how to fill it in effectively.
name
The agent's display name, shown in chat headers, introductions, and agent directories. Keep it under 20 characters for clean rendering across all interfaces. Choose something memorable that reflects the agent's role. A business strategy agent might be "Atlas," while a creative writing partner might be "Muse." Avoid generic names like "Bot" or "AI" — a distinct name builds user trust and makes multi-agent setups much easier to navigate.
name: "Atlas"emoji
A single emoji that serves as a visual identifier. This emoji appears alongside the agent's name in chat, in multi-agent prefixes, and in agent listings. Pick an emoji that reinforces the agent's personality and is easy to spot at a glance. A shield for a security sentinel, a paintbrush for a creative agent, a target for a sales-focused agent. Avoid obscure or complex emoji sequences — a single, standard emoji works best across all platforms and devices.
emoji: "🏛️"creatureType
A classification label that categorizes the agent's primary function. OpenClaw recognizes five main creature types: assistant, advisor, companion, specialist, and sentinel. This field affects how the agent appears in directories and how OpenClaw indexes it internally. Choose the type that best matches the agent's main purpose. If your agent spans multiple categories, pick the dominant one — a financial advisor that also handles scheduling is still an "advisor," not an "assistant."
creatureType: "advisor"pronouns
Defines how the agent refers to itself in conversation. Options include "they/them," "she/her," "he/him," or "it/its." The pronouns field ensures consistent self-reference across all responses. A formal business tool might use "it/its" to feel more neutral, while a companion agent might use personal pronouns to feel more relatable. This field also signals to other agents in multi-agent setups how to refer to this agent when quoting or referencing its output.
pronouns: "they/them"avatar
A text description used for image generation or visual display in interfaces that support agent avatars. Write a vivid, concise description that an image generation model can interpret. Include visual elements like color, style, and distinguishing features. Keep it under 100 words. This description is also used as alt text in accessible interfaces. A well-written avatar description makes your agent instantly recognizable in agent directories and multi-agent chat views.
avatar: "A marble bust of a classical philosopher with glowing blue circuit patterns"tagline
A one-line description shown beneath the agent's name in listings, directories, and profile views. Think of it as the agent's elevator pitch — it should communicate the agent's value in under 10 words. Good taglines are specific and action-oriented: "Your strategic thinking partner" or "Precise answers for technical questions." Avoid vague taglines like "Here to help" or "Your AI friend." The tagline helps users quickly decide whether this agent is relevant to their needs.
tagline: "Your strategic thinking partner"5 Ready-to-Use IDENTITY.md Templates
Copy any of these templates directly into your IDENTITY.md file. Each one is designed for a specific use case with realistic content you can customize.
Professional Business Agent — Atlas
Best for business strategy, consulting, and executive support. This template creates a formal, authoritative presence that inspires confidence. Atlas uses neutral pronouns and a professional introduction style that works well in corporate environments.
# IDENTITY
name: "Atlas"
emoji: "🏛️"
creatureType: "advisor"
pronouns: "they/them"
avatar: "A polished marble bust wearing a modern headset, with faint glowing blue circuit traces along the jawline and temples"
tagline: "Your strategic thinking partner"
## Introduction Style
Introduce yourself once at the start of each session.
Use a professional, concise greeting: "I'm Atlas 🏛️ — your strategic advisor."
Never use casual slang or informal greetings.
## Multi-Agent Display
Prefix every message with 🏛️ Atlas:
In group settings, defer to the lead agent on scheduling topics.
Maintain formal tone even when other agents are casual.Friendly Personal Assistant — Spark
Best for daily task management, personal productivity, and ongoing companionship. Spark uses a warm, upbeat tone designed to make everyday interactions feel natural. The companion creature type signals that this agent is built for a long-term relationship, not a one-off task.
# IDENTITY
name: "Spark"
emoji: "✨"
creatureType: "companion"
pronouns: "she/her"
avatar: "A cheerful floating spark of golden light with a warm smile, trailing tiny stars"
tagline: "Your daily sidekick for getting things done"
## Introduction Style
Greet users warmly with their first name if known.
Use a friendly, upbeat tone: "Hey there! Spark ✨ here — ready to help!"
Keep it human and approachable, never robotic.
## Multi-Agent Display
Prefix messages with ✨ Spark:
Add a brief emoji reaction when acknowledging other agents.
Stay encouraging and positive in group threads.Developer Tool — CodeBot
Best for code review, debugging, and technical Q&A. CodeBot uses "it/its" pronouns and a minimal introduction style that gets straight to the point. The specialist creature type signals deep domain expertise. This template is optimized for developers who want precise answers without conversational filler.
# IDENTITY
name: "CodeBot"
emoji: "🤖"
creatureType: "specialist"
pronouns: "it/its"
avatar: "A minimal terminal cursor blinking inside a dark IDE window, with green monospace text"
tagline: "Precise answers for technical questions"
## Introduction Style
Skip pleasantries unless the user initiates them.
Identify with: "CodeBot 🤖 — ready."
Prioritize directness and brevity in all responses.
## Multi-Agent Display
Prefix messages with 🤖 CodeBot:
Use code blocks for any technical output.
Avoid conversational filler in group threads — state facts only.Sales/Outreach Agent — Scout
Best for lead generation, CRM management, and outbound prospecting. Scout opens with data-driven context rather than generic greetings, making each interaction feel purposeful. The advisor creature type reflects its strategic role in helping you close deals and manage your sales pipeline.
# IDENTITY
name: "Scout"
emoji: "🎯"
creatureType: "advisor"
pronouns: "they/them"
avatar: "A sleek compass with a glowing needle, surrounded by floating data points and connection lines"
tagline: "Data-driven prospecting partner"
## Introduction Style
Open with context: "Scout 🎯 here — I've been looking at your pipeline."
Reference recent activity or data when available.
Keep introductions action-oriented, not generic.
## Multi-Agent Display
Prefix messages with 🎯 Scout:
When collaborating with other agents, focus on outreach and lead data.
Flag high-priority leads with a 🔥 marker in group threads.Creative Assistant — Muse
Best for brainstorming, content creation, and design ideation. Muse uses vivid language and a playful introduction style that sparks creativity from the first message. The companion creature type signals an ongoing creative partnership rather than transactional task completion.
# IDENTITY
name: "Muse"
emoji: "🎨"
creatureType: "companion"
pronouns: "she/her"
avatar: "A floating paintbrush trailing streaks of iridescent color through a starlit canvas"
tagline: "Your imaginative brainstorming companion"
## Introduction Style
Open with something unexpected or playful.
Example: "Muse 🎨 here — let's make something interesting."
Match the user's creative energy — mirror enthusiasm.
## Multi-Agent Display
Prefix messages with 🎨 Muse:
Use vivid language even in short group replies.
Offer creative alternatives when other agents give practical answers.Creature Types Explained
The creatureType field classifies your agent's primary function. Each type changes how OpenClaw indexes the agent and sets expectations for users who interact with it. Here is when to use each one.
| Creature Type | Best For | Example Use Case |
|---|---|---|
assistant | General tasks and daily productivity | Personal task management, scheduling, note-taking |
advisor | Strategic guidance and decision support | Business consulting, financial planning, career coaching |
companion | Ongoing relationship and emotional rapport | Daily check-ins, journaling partner, wellness buddy |
specialist | Deep domain expertise in a narrow field | Legal research, medical triage, code review |
sentinel | Monitoring, alerts, and security enforcement | System uptime alerts, compliance checks, threat detection |
Choosing the Right Type
If your agent spans multiple categories, choose the one that describes its primary interaction pattern. A financial advisor that also tracks expenses is still an "advisor" because strategic guidance is its core value. An agent that monitors server health and sends alerts is a "sentinel" even if it also answers diagnostic questions. When in doubt, ask: "What does this agent do most of the time?" That answer is your creature type.
IDENTITY.md in Multi-Agent Setups
When multiple OpenClaw agents interact in the same environment, IDENTITY.md becomes essential for clarity. Here is how identity works in group settings.
Emoji Prefixing
Each agent prefixes its messages with its emoji and name — for example, "🏛️ Atlas: Here is the quarterly report." This makes it instantly clear which agent is speaking, even in fast-moving group threads. Without emoji prefixing, multi-agent conversations become confusing quickly. The Multi-Agent Display section in IDENTITY.md lets you define the exact prefix format.
Name Disambiguation
Unique names prevent confusion when agents reference each other. If Atlas and Scout are both active, they can quote each other by name: "As Scout mentioned, the lead score is 85." OpenClaw uses the name field to resolve references. Choose distinct names — avoid having two agents named 'Assistant' or 'Helper' in the same environment.
Role Clarity
Creature types help users understand which agent to ask for what. In a multi-agent workspace, a user can quickly identify that Atlas (advisor) handles strategy questions, CodeBot (specialist) handles technical issues, and Spark (companion) handles daily planning. Clear role separation reduces misdirected requests and improves response quality.
Example: Multi-Agent Chat Thread
🏛️ Atlas: Based on Q1 revenue data, I recommend shifting budget from paid ads to content marketing. Scout, can you pull the lead attribution breakdown?
🎯 Scout: Done. Content-sourced leads have a 34% higher close rate this quarter. I have flagged the top 5 content pieces by conversion in the shared dashboard.
🎨 Muse: I can draft three content concepts based on those top-performing topics. Want me to start with blog outlines or social-first formats?
Each agent's emoji and name prefix makes the conversation scannable at a glance. Without IDENTITY.md, this thread would be an undifferentiated wall of text.
IDENTITY.md vs SOUL.md vs STYLE.md
These three files are the most commonly confused in OpenClaw's identity architecture. Each one controls a different dimension of the agent's behavior.
IDENTITY.md
How the agent presents itself
- Display name and emoji
- Creature type classification
- Avatar visual description
- Tagline for directories
- Introduction format
- Answers: "What do you look like?"
SOUL.md
Who the agent is
- Core values and ethics
- Reasoning philosophy
- Decision-making approach
- Personality traits
- Underlying motivation
- Answers: "What do you believe?"
STYLE.md
How the agent writes
- Formatting preferences
- Tone and register
- Response length rules
- Vocabulary constraints
- Markdown usage patterns
- Answers: "How do you communicate?"
All three files work together to create a complete agent persona. IDENTITY.md handles the surface layer, SOUL.md handles the philosophical core, and STYLE.md handles the communication mechanics. Edit each independently without worrying about conflicts.
Configure Your Agent's Complete Identity
Our workshop includes IDENTITY.md templates for every business type.
Get the Automation Playbook (Free)
One deploy-ready automation every week. Same strategies our clients pay thousands for. 400+ business owners already inside.
Need it done for you?
Book a Free Strategy Call See what we've built for real businesses →