Panofy

Getting Started with Panofy

On this page

Introduction

Panofy is the world's first agent training platform. With a single sentence, you can train a personalized agent that efficiently carries out research, generates materials, and even powers a knowledge Q&A bot — all based on your instructions. The core idea behind Panofy is simple: give everyone an agent with memory, personality, and the ability to continuously evolve — train once, and it works every time.

Most AI tools start from scratch every session. Every interaction requires you to re-establish context, re-explain your needs, and re-specify your output format. Panofy works differently — once those definitions are in place, the agent works to those standards consistently, learning from each interaction to understand you better.

The range of use cases is wide:

  • Content Assistant — Understands your brand voice and continuously generates on-tone copy and content ideas.
  • Knowledge Bot — Upload SOPs and product docs to create an onboarding guide or customer-facing support bot.
  • Analysis Reports — Define a competitive analysis framework and produce structured reports to the same standard, every time.

What these scenarios share is that the agent doesn't just answer once — it becomes a stable, long-term part of how you work.

Account Setup

Creating an Account

Sign in with your Google or GitHub account — no additional information required.

Login page with Google / GitHub one-click authorization
Login page with Google / GitHub one-click authorization

After logging in, you will be guided through an onboarding tutorial in the Agent Workspace.

Onboarding tutorial after login
Onboarding tutorial after login

Account Settings

Click the avatar in the bottom-left corner to open the account panel, where you can view your current subscription plan and available credits.

Click the bottom-left avatar to view subscription plan and credits
Click the bottom-left avatar to view subscription plan and credits

Activating an Invite Code

After registering on the homepage, you will automatically be added to the invite code waitlist. Check your email for your invite code.

Invite code received via email
Invite code received via email
Enter the invite code to activate your account
Enter the invite code to activate your account

Training Your Agent

Training is the core step in Panofy. Once training is complete, your agent retains the role definition, task scope, and output requirements you have set, and applies them consistently in every subsequent conversation — no need to re-explain the background each time.

Step 1: Create a new agent

In the Agent Workspace, click "+ New Agent" in the top-right corner to begin the creation process.

Agent Workspace homepage with the "+ New Agent" button highlighted
Agent Workspace homepage with the "+ New Agent" button highlighted

Step 2: Name your agent

Name your agent based on its intended function — for example, "TweetForge".

Enter agent name in the basic information field
Enter agent name in the basic information field

Step 3: Provide training instructions

In the training input field, provide instructions or knowledge content for your agent. Define its role, task scope, output format requirements, and tone preferences.

The more specific your training instructions, the more precise the agent's execution. Be sure to clarify: who it is, what it's good at, what format to output, and tone preferences.

Fill in role definition and task scope in the training input field
Fill in role definition and task scope in the training input field

Step 4: Upload reference materials

Upload files to accelerate knowledge building. Suitable materials include SOPs, business guidelines, evaluation frameworks, and reference examples.

Drag or click to upload training documents — supports PDF, Word, TXT, MD formats
Drag or click to upload training documents — supports PDF, Word, TXT, MD formats

Step 5: Submit training

Click "Submit" to begin training. You will be automatically redirected to the Agent Workspace, where training progress updates in real time.

Agent card showing "Training" status
Agent card showing "Training" status

Step 6: Training completes

Once training is complete, the agent's status changes from "Training" to "Ready". You can then proceed to the next step.

Assigning Tasks to Your Agent

Once training is complete, your agent has internalized its setup and is ready to receive and execute task instructions.

Executing a task

In the Agent Workspace, click on the relevant agent to open the conversation interface.

Click an agent to enter the conversation interface — files on the left, task list on the right
Click an agent to enter the conversation interface — files on the left, task list on the right

Enter your task instructions in the input box and send. The agent will begin executing the task automatically.

Send your task instructions
Send your task instructions
After sending instructions, the agent enters "Thinking" state
After sending instructions, the agent enters "Thinking" state

Results are displayed in real time in the conversation window and simultaneously saved to the "Output" panel on the left.

Task complete — results shown in the conversation, generated files in the left panel
Task complete — results shown in the conversation, generated files in the left panel
Generated promotional images
Generated promotional images

Creating a new task

Click "New Task" in the right-hand navigation bar to create an additional task for the same agent.

"New Task" button highlighted in the right navigation bar
"New Task" button highlighted in the right navigation bar

For complex tasks, the agent will automatically generate a task plan listing each step. You can then choose to execute immediately, schedule execution, or revise the plan.

Agent auto-generates a task plan — supports immediate / scheduled execution or replanning
Agent auto-generates a task plan — supports immediate / scheduled execution or replanning

Once execution is confirmed, you can monitor the progress of each sub-task in real time.

Right-side "In Progress" panel shows real-time sub-task progress
Right-side "In Progress" panel shows real-time sub-task progress

Click "Import" in the right-hand navigation bar to bring in an existing task plan from another complex task. This allows established workflows to be saved and reused directly.

Import an existing execution plan from other tasks to reuse established workflows
Import an existing execution plan from other tasks to reuse established workflows

Inviting Collaborators and Sharing

Click "Invite" in the right-hand navigation bar to generate an invite link. Anyone with the link can use your trained agent directly, with any credits consumed charged to their own account. You can configure whether to include access to external tools and whether to require a passcode.

Invite panel: control invite toggle, external tool permissions, passcode access
Invite panel: control invite toggle, external tool permissions, passcode access

Click the share icon above the conversation to generate a shareable link for the task record. You can set visibility to "Anyone with the link" or "Only me", and optionally enable playback mode so external viewers can replay the task without needing to register.

Click the share icon in the top navigation bar (highlighted)
Click the share icon in the top navigation bar (highlighted)
Set sharing permissions — enable playback mode for external viewers without registration
Set sharing permissions — enable playback mode for external viewers without registration