VoyagerAvtonomni agent z LLM, ki se uči in raziskuje v Minecraftu brez človeškega vnosanja.

4.8 (5)

Pregledal Daniel Nikulshyn·Posodobljeno maj 2026

Open Source Research LLM Agents Autonomous Agents Game AI Reinforcement Learning Minecraft

Pregled

Voyager je raziskovalni projekt, ki uporablja LLM za upravljanje avtonomnega agenta v Minecraftu. Agent si postavlja lastne cilje, napiše izpelnljivo kodo, da deluje v svetu, in postopoma gradi knjižnico ponovno uporabnih spretnosti, medtem ko igra. Združuje samodejni učni načrt za odprto raziskovanje, iterativno zanko poizvedb, ki kodo izboljšuje na podlagi povratnih informacij okolja, in naraščajočo zbirko spretnosti, ki agentu omogoča reševanje vedno zahtevnejših nalog. Sčasoma Voyager odklanja nove mejnike tehnološkega drevesa, zbira različne predmete in prehaja več terena kot prejšnji Minecraft agenti. Voyager je predvsem zanimiv za raziskovalce umetne inteligence, razvijalce AI v igrah in hobištence, ki raziskujejo embodied agents, stalno učenje in odločanje na podlagi LLM v okoljih z odprtim svetom.

Ključne funkcije

Avtomatski načrt za generiranje ciljev
Iterativno povpraševanje z povratnimi informacijami iz okolja
Rastoča zbirka spretnosti z izvedljivo kodo
Načrtovanje in razmišljanje z LLM
Raziskovanje brez konca v Minecraftu
Raziskovalno usmerjena, odprtokodna implementacija

Cene

Model: Free
Kategorija: Igre
Ocena: 4.8 / 5 (5)

Primeri uporabe

Benchmark LLM agentov v Minecraftu

Raziskovalci lahko ocenijo avtonomne agente z LLM na nalogah z odprtim koncem v Minecraftu, pri čemer primerjajo napredek tehnološkega drevesa, raznolikost predmetov in raziskovanje z prejšnjimi merili.

“""”

Študij celoživotnega pridobivanja spretnosti

Uporabite rastočo zbirko spretnosti Voyagerja in avtomatski načrt, da preučite, kako agenti nabirajo ponovno uporabljive kode spretnosti na dolgi rok brez človeškega nadzora.

“""”

Prototipiranje AI vedenj v igrah

Razvijalci AI v igrah lahko eksperimentirajo z načrtovanjem z LLM in iterativnim izboljševanjem kode, da ustvarijo avtonomne NPC-je, ki si postavljajo cilje in se prilagajajo z povratnimi informacijami iz okolja.

“""”

Praktično učenje za hobije

Hobiji, ki raziskujejo LLM agente, lahko poženijo Voyager, da vidijo transparentne, pregledne akcije kode in se naučijo, kako zanke poizvedovanja in načrti vodijo raziskovanje brez konca.

“""”

Prednosti in slabosti

Prednosti

Raziskovanje brez konca, celoživotno učenje brez človeškega posredovanja
Vzpostavlja ponovno uporabljivo zbirko spretnosti, ki se z časom povečuje
Odlično benchmark rezultate v primerjavi z prejšnjimi agenti Minecrafta
Transparentne, na kodi podprte akcije so enostavne za pregled

Slabosti

Potreba po dostopu do zmogljive LLM API, kar je lahko drago
Omejeno na Minecraft kot okolje
Nastavitev in prilagajanje lahko zahtevata napredne tehnične zahteve
Delovanje močno odvisno od kakovosti vnaprejšnjih vprašanj in modela

Ocene

4.8

Povprečje iz 5 ocen.

Prijavi se za oddajo ocene.

Sofia Lindqvist

Apr 26, 2026

Use it every day

Honestly didn't expect to like it this much. Growing skill library of executable code is exactly what I needed, and builds a reusable skill library that compounds over time. I do wish performance depends heavily on prompt and model quality, but I reach for it almost every day now and it just clicks.

Marcus Bell

Nov 18, 2025

Skeptical, then convinced

I went in skeptical — most tools in this space overpromise. It actually delivers on iterative prompting with environment feedback, and open-ended, lifelong learning without human intervention caught me off guard. Performance depends heavily on prompt and model quality is why this isn't a perfect score, still, I'd recommend giving it a real trial.

Liam O’Connor

Oct 7, 2025

Use it every day

Honestly didn't expect to like it this much. Iterative prompting with environment feedback is exactly what I needed, and strong benchmark performance versus prior Minecraft agents. but I reach for it almost every day now and it just clicks.

Nadia Petrova

Sep 17, 2025

Compared a few options

Evaluated this against two competitors. Where it wins: iterative prompting with environment feedback and builds a reusable skill library that compounds over time. On balance the feature set — especially automatic curriculum for goal generation — justifies the 5 stars for our use case.

Aisha Khan

Sep 6, 2025

Solid for our team

We rolled this out across the team last quarter and open-ended, lifelong learning without human intervention. Automatic curriculum for goal generation fits neatly into how we already work, and iterative prompting with environment feedback removed a step we used to do by hand. Performance depends heavily on prompt and model quality, which is the main caveat, but it has held up under daily use.

Vprašanja

What are the main limitations of Voyager's autonomous agent?

Performance depends heavily on the quality and cost of the LLM, it is confined to Minecraft as the only environment, and the open-ended curriculum may require careful tuning to avoid undesired behaviors.

Asked by Tomáš Novák · Dec 20, 2025

Who can realistically use Voyager and what technical skill is required?

The tool targets AI researchers, game AI developers, and hobbyists; users need to be comfortable setting up Python environments, configuring LLM API keys, and tuning prompts, as the setup is technically involved.

Asked by Winifred Adeyemi · Nov 30, 2025

Which Minecraft versions or platforms does Voyager support?

Voyager is built for the standard Java Edition of Minecraft and interacts with the game via its existing modding or server APIs; it does not currently support Bedrock Edition or other game engines.

Asked by Olga Ivanova · Nov 19, 2025

What are the costs associated with running Voyager?

Voyager itself is open-source, but it relies on a capable LLM API (e.g., OpenAI, Anthropic) for planning, and those API calls incur usage fees that can become significant depending on runtime and model size.

Asked by Marcus Bell · Oct 26, 2025