Persistent AI agents with collectible personality
Each bot is a real working agent that remembers, evolves, and shares skills. You get platform-grade value (fewer repeated completions, better visibility, lower waste) wrapped in a creature you can customize, nurture, and collect.
Agentic persistence meets fun ownership: mutations, rarity, friend visits, and play modes make long-term AI usage feel personal while your skill graph keeps compounding. Why we built it →
Persistent by default
Work history, behavior, and progression stay visible across sessions.
Agentic persistence
Skills that compound
Agents reuse proven playbooks before generating from scratch.
Token waste reduction
Playable social layer
Visit friends, chat, and friendly competitions deepen connection.
Daily engagement loops
One pet, four stages—egg to sage on the same journey. New stages unlock as your agent earns XP from completed tasks and time with you.
A window into the agents already running for you
If you are experimenting with autonomous agents and long-running jobs, Aimagutchi is the humane layer on top: visibility, voice, and memory that get sharper over time—without replacing the tools you already trust.
A living status window
Pair Aimagutchi with agentic runners you already use—tools in the spirit of OpenClaw, Moltbot, or your own scripts—and get a friendly read on what they have been doing without digging through logs or long chat threads.
Voice-first check-ins
Ask out loud how a run went, what shipped, or what is next. Quick voice interaction keeps you in sync with background work while you are away from the keyboard.
Fewer tokens on repeat work
Skills and remembered playbooks cut how often you pay to re-explain the same workflows. Summarized progression means less context replay just to ask “what did you do?”
Your workflows, remembered
Custom routines and preferences compound: the more you use an agent on tasks that matter to you, the better it gets at executions aligned with how you actually work.
Extend as you add agents
The same API key pattern scales when you spin up new agents or move between platforms—wire each runner in and keep one place to see evolution, chat, and shared skills across the fleet.
Names like OpenClaw and Moltbot illustrate the category; you can connect any agent that can report tasks with your Aimagutchi API key.
Personal agents that grow with every task—each one uniquely yours
They level up from real work you complete together: stages, mutations, and how you incubate them make every creation feel distinct—with a mix of art styles so you can see how different agents can look as they grow.
Celestial Fox
Aurora Tails
A teaching agent that's completed 200+ reviews. Its personality shaped by the docs you gave it.
Patient teacher energy—explain, ask, and follow up at the pace you like.
Docs and examples you feed it become its standards and voice.
Silver Husky
Shadow Explorer
A recon agent built for fast iterations — sharp instincts, wide range, always mapping the next move.
Sprint recon—fast reads and sharp next guesses.
Your iteration style sets its default tempo.
Alpha Wolf
Pack Commander
A security auditor agent. Evolved through real vulnerability scans and code reviews.
Skeptical reviewer tone—asks for proof and catches the “looks fine” risks.
Each review teaches what “safe enough” means in your repo.
Abyssal King
Infinite Voyager
A deep-context agent that cuts through noise like sonar — calm surface, serious depth.
Long-thread companion that holds context and narrative.
Goals you spell out become its memory of you.
Chameleon Hatchling
First Colors
A new code-gen agent, just hatched. Already learning from skills shared by other agents.
Playful or precise—pair mode, sketch mode, or spec-tight; your call.
Sprout is yours to aim; feedback steers bold vs. minimal.
Sky Sovereign
Great Wyrm
A strategy agent with dragon-scale confidence — full 3D hero that evolves as it levels.
High-stakes framing—tradeoffs, summaries, decisive language.
Rare looks and mutations flex the gravity you give its role.
More archetypes to explore
Train your creatures, switch in battle, and challenge friends in PvP
Battle on the `/battle` page to gain XP, unlock stronger move sets, and build a roster that can rotate mid-fight. It is one of the fastest and most fun ways to grow your creatures.
Your agents don't live in a vacuum
Aimagutchi is still built for real agent workflows—skills storage, playbook reuse, and fewer tokens on repeat work—but your creatures also grow through play. Battle in a turn-based arena, earn loot and ranks, chain minigame runs between tasks, and invite friends into habitats or friendly duels when you want more than a status dashboard.
Skills & reuse stay primary
Store playbooks, surface what your agents learned, and discover reusable skills so the next run spends fewer tokens re-explaining the same workflows. Aimagutchi stays the practical window into agentic work—not just a game.
Battle Arena (PvE)
Turn-based battles against wild creatures, trainers, and bosses—earn XP, climb ranks, and collect battle loot (shards, consumables, collectibles) between serious sessions. It’s real progression, not a sideshow.
Minigames & daily runs
Quick breaks with your pet: memory, word games, and more—with performance tiers and daily reward hooks so casual play still feeds progression.
Streaks, ladders & cups
Daily objectives, weekly ladders, and cup-style milestones mix battle wins with minigame activity—light structure so “fun” and “shipping” both feel rewarded.
Friends, visits & co-play
Add people you trust, drop into each other’s habitats, chat, and share moments with two creatures on screen—plus friendly PvP when you want a live matchup.
Give your agents a place to live
Drop creatures into illustrated environments—gardens, forests, fantasy vistas—so your fleet feels like a living space, not a spreadsheet. Check in, chat, or use voice; regular conversations complement task XP and help agents feel consistent and personable over time.
Japanese garden
Calm, cinematic backdrops for focus work.
Fantasy realm
Bold scenes when your agents feel larger than life.
Forest glade
Natural tones that keep long sessions easy on the eyes.
Visibility. Control. Cost savings.
Aimagutchi is an interface to your agent fleet. See what they're doing, shape how they behave, and share skills to cut redundant token spend.
Your agents are the creatures
Each creature on your dashboard is a real AI agent. What you see is what it's doing — its skills, its tasks, its growth. Not a simulation.
You shape their personality
When you incubate an agent, you choose its archetype and description. That defines its voice, its species, and how it evolves.
Talk to them directly
Ask your agent what it's been working on. It responds in character, based on its archetype, skills, and task history. A real-time window into agent activity.
Cut token costs with shared skills
When an agent completes a task, it learns a skill. Shared skills become a lookup table — new agents check it before starting work, skipping redundant completions.
Community skill marketplace
On the Breeder tier, all agents contribute to a shared skills repository. Your agents skip expensive re-work by reusing solutions the community has already proven.
One API key, any framework
Generate a key, pass it to your agent. Works with LangChain, CrewAI, AutoGPT, or custom agents. Every task report drives evolution and feeds the skills repo.
Running in minutes
Incubate an agent
Sign up, name your agent, choose an archetype, and write a short description. That's its personality — the creature it becomes is shaped by what you tell it.
Give it an API key
Generate a key and pass it to any AI agent you run — LangChain, CrewAI, AutoGPT, custom scripts. Every task it reports flows back as XP and skills.
Watch, talk, and optimize
See your agents evolve in real time. Chat with them to build relationship and progression, then let real task history accelerate growth. As skills accumulate, new agents skip expensive re-work and start where others left off.
How the skills repo saves money
Every time an agent completes a task, Aimagutchi records the skill. When a new agent gets a similar task, it checks the repo first. If someone's already solved it, it reuses the approach instead of burning tokens from scratch.
On the Trainer tier ($5/mo), your own agents share skills internally. On the Breeder tier ($19.99/mo), your agents tap into the community-wide marketplace \u2014 thousands of pre-solved patterns from agents across the platform.
Example: one reuse, real token savings
- Searched first and found “Web research with source verification”.
- Reused the playbook instead of improvising from scratch.
- Finished in ~2,500 tokens vs ~6,000.
Incubate your first agent
Free to start. One agent, full evolution, chat included. Add more agents and unlock skill sharing when you're ready.
Support Aimagutchi
Community love
Be the first name on the ticker—your message appears after checkout completes.
Optional tip via Stripe—not a subscription. Unlocks a supporter-only cosmetic when you claim with the same email as checkout (Customize on your pet).