PERSONAL AI / AGENT IDENTITY
Give Your Agent A Consistent Role
Define how your agent should help, what you want it to remember and where your preferences apply. Keep that context in a base it can return to across conversations, with lessons you can inspect and revise.
OUR ASSISTANT STORY
One Assistant, Different Kinds of Help
Our assistant works with us on different tasks. We built Agent Identity to give it a clear role while leaving room for the kind of help each situation needs.
WHAT HELPFUL MEANS
Sometimes We Need a Question
A quick update needs a clear next action. A design discussion may need questions, alternatives and tradeoffs. We wanted our assistant to understand the kind of help we were asking for, including when to challenge an assumption or ask before acting.
WHAT EXPERIENCE TEACHES
A Correction Needs Its Context
A suggestion can be useful in one situation and wrong in another. We needed to keep what happened, what we learned and where that lesson applies. A correction should help the next relevant task without turning every exchange into a lasting preference.
AN IDENTITY YOU CAN SHAPE
A Role You Define, Experience You Can Review
Agent Identity gives the assistant a place for its role, the way you want to work together and lessons supported by experience. You can ask it to remember a preference, revisit a correction or review a proposed lesson from an earlier task.
The included startup instruction brings the relevant parts into the next conversation. Current work stays in its project base; the identity holds the context for how the assistant approaches that work.
Video coming soon
We’re preparing a walkthrough for this section.
WHY IT WORKS
A Clear Role With Room to Adapt
Carry a useful role between sessions, apply preferences where they belong and build on reviewed experience.
CONTINUITY BETWEEN SESSIONS
Return to a Familiar Role
A new session can start from the role you have already defined.
Your startup instruction points the agent to its saved role, boundaries and preferences. The agent can apply that context to the next task and conversation.
PREFERENCES WITH CONTEXT
Adapt to the Situation
A progress update and a detailed answer need different styles.
The agent checks where your preferences apply before using them in the current task. A request to keep reports brief still allows detailed explanations.
REVIEWED EXPERIENCE
Build on What You Learn
Useful experience can inform the next task after review.
The agent saves preferences when you ask it to remember them. Review proposed lessons with their evidence and limits before using them in future tasks.
TRY IT YOURSELF
Start with One Small Task
Try Agent Identity on a guided exercise before defining your agent’s role and preferences. Need to install Lorebo or connect your agent? Get Started
Find Agent Identity in Templates → Personal AI in Lorebo Desktop. Download it and apply it to a writable base, then open Start Here for your first task.
Video coming soon
We’re preparing a walkthrough for this section.
Ask your agent to follow Agent Startup Instruction and try a Practice Library exercise. Inspect the result and which guidance it used; the trial saves nothing.
Video coming soon
We’re preparing a walkthrough for this section.
Save your agent’s role and one preference. Add the included portable opening instruction to your chosen client, then test a new session and a different task.
Video coming soon
We’re preparing a walkthrough for this section.
FREQUENTLY ASKED QUESTIONS
Good to Know
What’s included in the template?
A startup instruction, reusable workflows, record templates and three guided exercises. They help you define your agent’s role, save preferences with their context and review lessons. The starter contains no personal profile or past conversations.
How does it differ from a system prompt?
A short startup instruction tells the agent when to consult the base. Its role, preferences and supporting experience live in connected pages that you can maintain separately. The agent retrieves what matters for the current task.
How does my agent start using it?
Select your writable copy and ask the agent to read Agent Startup Instruction. For future sessions, use its portable opening instruction in your chosen client. Importing the template does not install a skill or configure that client.
Does every chat become a memory?
No. Your agent saves preferences when you ask it to remember them. Proposed lessons from experience are reviewed before guiding future work.
Does activating the template make it writable?
Activate only opens the template in read-only mode. To use it fully, create a base from it as shown in Try It Yourself
Can I start in Free?
Yes. In Free, check that your active local base is empty and writable, then choose Replace Active Base. Lorebo makes a backup before replacing it. Pro also offers Register and Merge.
Can it work with my other bases?
Yes, when your agent has access. Identity supplies its role, preferences and reviewed lessons; a course or project base holds its own facts and current tasks. Identity does not grant additional permissions.
Will the agent always follow it?
The base provides guidance you can inspect; it does not guarantee every response. Ask which guidance was used and check the result on a small task. Current instructions and the agent’s permissions still apply.