BLOGOfficialAgent 101: Launching production-grade agents at scaleNebius walks through the four pillars of production AI agents (LLMs, agent frameworks, evaluation, memory) with hands-on integration examples for CrewAI, Google ADK, LangChain and Agno on Token Factory.
REPOOfficialAgentic workflows with OpenClaw on ServerlessRun agent-style pipelines with tool use, retrieval, and multi-step reasoning using OpenClaw on Serverless GPU compute.
VIDEOOfficialAGENTS.md Is Making Your Agent WorseCommon pitfalls in agent prompt engineering — what AGENTS.md gets wrong and what to do instead.
VIDEOOfficialAgent Testing: Browser and CLIEnd-to-end testing strategies for browser and CLI agents. Practical patterns for keeping agents reliable.
DOCSOfficialAgno integrationLightweight multi-modal agent framework with Token Factory as a first-class provider.
VIDEOOfficialAI Day: Nemotron — Open and Accelerated (Bryan Catanzaro, NVIDIA)AI Day by Nebius Academy. Bryan Catanzaro on NVIDIA's open-model approach with Nemotron and the future of accelerated AI infrastructure.
VIDEOOfficialAI DNA Amsterdam — RecapTalks, discussions, and connections from the AI DNA event in Amsterdam. Highlight reel.
DOCSOfficialaisuite integrationUse Token Factory as a model backend in aisuite, the model-agnostic Python SDK from Andrew Ng.
BLOGOfficialAnnouncing the Tavily Web Search API certificationTavily's free certification program teaching developers to build agents that search, validate and explore the web, covering search fundamentals plus the Extract and Crawl tools.
VIDEOOfficialAutomated Documentation: Backend & FrontendPractical patterns for keeping docs in sync with code, backend and frontend. Tooling + agent flows.
DOCSOfficialAutoscaling and cache-aware routingHow Token Factory scales inference workloads and routes requests based on cache locality. Rate limits, burst behavior, and tuning guidance.
BLOGOfficialBare-metal-class performance for AI inference in MLPerf Inference v5.1Nebius proves near-bare-metal inference performance across three AI systems in MLPerf Inference v5.1, showing how its virtualized GPU stack serves production LLM inference without a meaningful performance tax.
DOCSOfficialBatch inferenceSubmit large request batches asynchronously for offline workloads. Cheaper per-token than real-time inference; ideal for evals and back-fills.
BLOGOfficialBeyond prompting: fine-tuning LLMs with NebiusA practical fine-tuning walkthrough covering dataset prep, LoRA adapter training on Llama-3.1-8B, deployment and evaluation against the Berkeley Function Calling Leaderboard on Nebius AI Studio (Token Factory).
BLOGOfficialBuild a multi-agent customer support system with Google ADK, A2A and NebiusStep-by-step guide to building an orchestrated support system where specialized agents handle sentiment, retrieval and escalation, powered by Nebius AI Studio (Token Factory) models, Google ADK and the A2A protocol.
DOCSOfficialBuild a RAG chatbot on Serverless AIEnd-to-end RAG: JupyterLab for prototyping, Serverless AI for inference, and Managed PostgreSQL for the vector store.
BLOGOfficialBuild a RAG-powered content generation platform with Nebius and QdrantBuilds a full RAG pipeline: chunking, Qwen3-Embedding-8B vectors stored in Qdrant Cloud, semantic retrieval, and personalized generation with Llama-3.3-70B served on Nebius AI (Token Factory).
BLOGOfficialBuilding an AI-Powered Finance Planner with Full-Stack Next.js and NebiusStep-by-step build of Money-Guard, a Next.js dashboard that analyzes spending and answers questions about transactions using Meta Llama 3.1 70B served by Nebius AI Studio.
WORKSHOPOfficialBuilding More Secure Autonomous AI Agents with NVIDIA NemoClawNemoClaw is an open-source blueprint for safer autonomous agents — it wraps OpenClaw with a kernel-level sandbox, an L7 egress proxy, and an intent-classifying policy engine. This demo-focused webinar shows two ways to run NemoClaw on Nebius, including via Token Factory.
PLAYLISTOfficialBuild with Token Factory playlist on YouTubeAI projects, tools, and success stories created with Token Factory — see what other builders are shipping.
VIDEOOfficialClaude Code Subagents TutorialParallel execution and custom agent patterns with Claude Code. Worked examples + best practices.
BLOGOfficialClusters vs single nodes: which to use for training and inferenceWhen to fine-tune on one machine versus scaling across many nodes — how PyTorch DDP, DeepSpeed and Horovod handle data partitioning and gradient sync, with Kubernetes/Slurm scheduling, illustrated on Nebius GPU infrastructure.
DOCSOfficialCrawl to RAG: turn any website into a searchable knowledge baseOfficial Tavily docs example that crawls and vectorizes a website with Tavily Crawl and MongoDB, then queries it through a conversational RAG agent.
BLOGOfficialCreate Your Own AI-Powered Code Generator and ReviewerBuild a full-stack Next.js code assistant that generates snippets across languages and returns automated reviews, powered by DeepSeek Coder on Nebius Token Factory.
DOCSOfficialCrewAI integrationBuild multi-agent crews with Token Factory models. Coordination, role-playing, and task pipelines all wired through Nebius inference.
BLOGOfficialData preparation techniquesBest practices and practical pipelines for preparing large datasets for LLM training and distributed AI workloads on GPU clusters.
DOCSOfficialDedicated endpointsPin a Token Factory model to dedicated GPU capacity for predictable latency at scale. Includes autoscaling rules and routing patterns.
VIDEOOfficialDefining Cursor's RulesGet consistent, predictable behavior from your AI pair programmer. How to author Cursor rules that hold.
DOCSOfficialDeploy a model as a Serverless EndpointStep-by-step tutorial for packaging a model, building the container, and exposing it via a Serverless Endpoint.
DOCSOfficialDeploy custom modelsTake a model you trained anywhere and deploy it behind a Token Factory endpoint. Custom weight loading, scaling, and monitoring.
WORKSHOPOfficialDeploying a Knowledge-Based Chatbot with RAG in ProductionNebius webinar with CSA Boris Popov on shipping a production RAG chatbot on Nebius AI Cloud, powered by NVIDIA H100 GPUs with Kubernetes, Triton, TensorRT, Milvus, and PyTorch.
DOCSOfficialDeploy your first Serverless EndpointStand up a long-running inference endpoint with a public URL. Token Factory-style API, custom container.
BLOGOfficialEffortless web-based RAG evaluation with Tavily and LangGraphTavily tutorial that combines its Search API with LangGraph to build an agent generating dynamic, subject-specific RAG evaluation datasets straight from live web data.
VIDEOOfficialEnglish = New Programming LanguageThe future of coding agents and natural-language software development. Where the agent paradigm is heading.
BLOGOfficialFault-tolerant training: how we build reliable clusters for distributed AINebius's multi-layered approach to reliable large-scale training — liveness probes, automatic checkpoint-restart on hardware failure, graceful node termination, and the MTBF/MTTR metrics behind a dependable GPU cluster.
DOCSOfficialFine-tuning on Serverless JobsUse Serverless Jobs to run fine-tuning workloads — data in, model artifact out — without managing the cluster.
DOCSOfficialFunction calling and tools with Token FactoryOfficial Token Factory guide on defining tools, letting models pick functions from context (or forcing a specific call), with Python and JavaScript examples for building tool-using agents.
BLOGOfficialFundamentals of LoRA and low-rank fine-tuningA Nebius deep dive into how LoRA updates a low-rank decomposition of weight matrices (W' = BA) to match full fine-tuning at a fraction of the cost, plus newer variants like PiSSA and DoRA.
VIDEOOfficialHow to Create Object Storage Buckets on NebiusQuick walkthrough for creating Object Storage buckets on Nebius — fundamentals you'll use for every training and inference job.
VIDEOOfficialHow to Deploy a Model with Nebius Serverless EndpointsFrom container to live API. Walks through deploying a model as a serverless endpoint on Nebius.
VIDEOOfficialHow to Fine-tune a Model with Nebius Serverless JobsEnd-to-end fine-tuning workflow using Nebius Serverless Jobs — from data prep to deployed checkpoint.
VIDEOOfficialHow to run Boltz-2 at scale on KubernetesDeploy a Kubernetes cluster, configure shared storage, and run reproducible, multi-node Boltz-2 inference for protein folding.
VIDEOOfficialHow to Run Boltz-2 at Scale on KubernetesEnd-to-end tutorial on running Boltz-2 protein structure inference at scale on Nebius MK8s. Pairs with the boltz2-mk8s repo.
BLOGOfficialHow to Run Meta Llama 3.1 405B with the Nebius AI Studio APIA hands-on how-to for calling Meta Llama 3.1 405B through the Nebius OpenAI-compatible API, with working Python, JavaScript, and cURL examples.
BLOGOfficialInference guide with vLLMDeploy high-throughput LLM inference on Nebius GPU cloud using vLLM and Kubernetes. Includes configuration patterns for scalable model serving.
BLOGOfficialIntroducing Managed Soperator: your quick access to Slurm trainingNebius's fully managed Slurm-on-Kubernetes solution is now self-service: spin up a training-ready GPU cluster with pre-installed drivers and libraries in minutes. Explains the Soperator operator, shared root filesystem, and automatic node recovery.
BLOGOfficialIntroduction to model distillation: efficient knowledge transferAn end-to-end tutorial that distills a Qwen3-235B teacher into a Qwen3-4B LoRA student on Nebius Token Factory, covering batch-inference data generation, fine-tuning, deployment and evaluation with a downloadable notebook.
REPOOfficialK8s on NebiusDeploy GPU-enabled Kubernetes clusters with pre-installed NVIDIA drivers and optimized networking for AI training and inference workloads.
REPOOfficialKvax: fast Flash Attention for JAX with context parallelismNebius's open-source Flash Attention 2 implementation for JAX, built on Triton kernels with efficient document-mask computation and context parallelism for FSDP/HSDP-sharded long-sequence training on GPU clusters.
DOCSOfficialLangChain integrationUse Token Factory chat models, embeddings, and retrievers inside LangChain via the langchain-nebius package.
DOCSOfficialLaunch your first Serverless JobRun a containerized AI workload to completion on Nebius GPUs without provisioning a cluster.
REPOOfficialLife sciences workloads on Serverless (OpenMM)Run molecular simulations and generate datasets on GPU. Recipes for OpenMM and friends, packaged for Serverless Jobs.
DOCSOfficialLiteLLM integrationRoute LiteLLM through Token Factory as an OpenAI-compatible provider. Drop-in for projects already using the LiteLLM proxy.
DOCSOfficialLlamaIndex integrationWire Token Factory in as the inference layer for LlamaIndex RAG pipelines.
REPOOfficialLLM inference — vLLM endpointServe large language models as real-time APIs using vLLM containers on Serverless GPU endpoints.
WORKSHOPOfficialMake the Most of Cursor as an AI-Powered Coding EnvironmentFree short video course featuring Cursor's co-founder + Nebius Academy experts. Learn agentic development end-to-end: context management, parallel agents, browser/CLI testing, custom rules, automated docs, and prompt patterns that hold up in production.
DOCSOfficialManaged Kubernetes docsReference documentation for Nebius Managed Kubernetes — cluster lifecycle, node groups, autoscaling, and GPU configuration.
VIDEOOfficialManaged Soperator: Slurm on KubernetesMarouane Khoukh introduces Managed Soperator — Slurm on Kubernetes for scalable AI workloads on Nebius.
REPOOfficialML Cookbook · DeepEPDeepEP (DeepSpeed Expert Parallelism) on Nebius. Configs and example launches for MoE training at scale.
REPOOfficialML Cookbook: pre-training DeepSeek-V3 with MXFP8 on a B200 clusterA Nebius ml-cookbook recipe with Slurm job scripts for multi-node pre-training of DeepSeek-V3 (16B and 671B) on a 256-GPU NVIDIA B200 cluster, showing up to 41% faster throughput with MXFP8 mixed precision and DeepEP.
REPOOfficialML Cookbook · Run:ai on NebiusRun:ai-based GPU orchestration on Nebius. Setup, scheduler config, and multi-tenant patterns.
REPOOfficialML Cookbook · SkyPilot on NebiusRun multi-cloud ML workloads on Nebius using SkyPilot. Recipe with sample configs and run scripts.
REPOOfficialML Cookbook · Slurm on NebiusSlurm-based training on Nebius compute clusters. Job submission scripts, sbatch templates, and orchestration patterns.
REPOOfficialML Cookbook · Volcano on Nebius MK8sVolcano scheduler on Nebius Managed Kubernetes for batch ML workloads. Manifests + queue configurations.
BLOGOfficialMLPerf Training v5.1: leading results on NVIDIA Blackwell and Blackwell UltraNebius posted seven first-place finishes in MLPerf Training v5.1 across NVIDIA Blackwell and Blackwell Ultra systems — a look at the hardware-software co-optimization behind training performance on Nebius AI Cloud.
VIDEOOfficialNebius AI Cloud Console overviewLearn how to provision GPU infrastructure, manage Kubernetes clusters, launch Slurm jobs, and monitor workloads from the console.
WORKSHOPOfficialNebius AI Cloud Onboarding CourseFree, self-paced course. Get from your first login to real cloud workflows in 30 minutes — navigate the console, launch storage and VMs, deploy and run an LLM, configure access for a dev environment.
REPOOfficialNebius Kubernetes Applications: Helm charts for GPU & AI workloadsOfficial Nebius repo of 50+ ready-to-deploy Helm charts and manifests for Managed Kubernetes — including vLLM inference, Ray cluster/serve, NVIDIA GPU Operator, Stable Diffusion WebUI, MLflow and JupyterHub. A practical reference for deploying AI apps on Nebius mk8s.
WORKSHOPOfficialNebius ScienceFollow how AI and science co-evolve through the lens of the cloud powering it. Research, interviews, and case studies on AI's impact on scientific discovery.
REPOOfficialNebius Solutions LibraryReference architectures and ready-to-deploy recipes for GPU training, inference, K8s, Slurm, and SkyPilot on Nebius AI Cloud.
PLAYLISTOfficialOfficial Token Factory playlist on YouTubeLearn about Nebius Token Factory from the makers — architecture, deep dives, and live demos curated by the team.
BLOGOfficialOrchestrating LLM fine-tuning on K8s with SkyPilot and MLflowManage distributed fine-tuning workloads with experiment tracking on Kubernetes GPU clusters.
REPOOfficialPost-training cookbook examplesEnd-to-end fine-tuning recipes for the Token Factory post-training stack — SFT, DPO, evaluation, and deployment.
DOCSOfficialPost-training guideFine-tune open models on Token Factory: data prep, training runs, evaluation, and deployment back to the inference API.
BLOGOfficialPost-training in Token Factory: the missing layer between MVP and productionExplains how Nebius Token Factory turns base models into production systems via data ingestion, multi-node supervised fine-tuning, reinforcement fine-tuning, speculative decoding and deployment.
DOCSOfficialPydantic AI integrationType-safe agent framework with Pydantic validation. Token Factory backs the inference layer.
VIDEOOfficialRecap & Next Steps with Nebius CloudRecap of what you've learned about Nebius Cloud and a roadmap for what to explore next.
BLOGOfficialRouting in LLM inference is the difference between scaling and stallingWhy request routing is the single biggest lever in production LLM inference, and how Token Factory routes intelligently.
BLOGOfficialRunning Boltz-2 inference at scale in Nebius AI CloudScale life-sciences inference workloads on GPU clusters optimized for performance and throughput.
BLOGOfficialRunning NVIDIA NIM and NVIDIA Blueprint in Nebius AI CloudDeploy validated healthcare AI models using NVIDIA NIM on high-performance GPU infrastructure.
VIDEOOfficialRunning OpenClaw on NebiusFrom zero to a live AI agent in 15 minutes. Walks through Token Factory, three deploy methods (npm / Docker / Nebius Serverless), and a hands-on agent build.
REPOOfficialServerless AI cookbookTop-level repo for every Serverless AI recipe — jobs, endpoints, RAG, agents, life sciences, voice, and more.
VIDEOOfficialServerless AI — full walkthroughDeep-dive video walkthrough of the Serverless AI model — when to use Jobs vs Endpoints, with real demos.
DOCSOfficialServerless AI overviewHigh-level overview of the two Serverless AI execution models — Jobs (run to completion) and Endpoints (real-time APIs).
REPOOfficialServerless Endpoints — first-endpoint examplesReady-to-deploy example endpoints from the Serverless cookbook. Build, push, and serve.
VIDEOOfficialServerless Endpoints — quickstart videoFive-minute walkthrough of deploying a Serverless Endpoint and hitting it from your app.
REPOOfficialServerless Jobs — first-job examplesReady-to-run example workloads for the Serverless Jobs quickstart. Clone, edit, and submit.
VIDEOOfficialServerless Jobs — quickstart videoFive-minute walkthrough of launching your first Serverless Job. Container in, GPU compute out.
VIDEOOfficialSharing Rules Between AgentsKeep coding agents consistent across teams and codebases. Patterns for shared rules and config.
REPOOfficialSkyPilot — finetuning & orchestratingOrchestrate multi-node LLM fine-tuning and distributed training across GPU VMs and Kubernetes clusters using SkyPilot.
DOCSOfficialSkyPilot on Nebius docsOfficial docs for the SkyPilot integration on Nebius AI Cloud. Provisioning, multi-node training, and autoscaling patterns.
DOCSOfficialSoperator docs (Managed Slurm)Operator-style API for running Slurm clusters on Kubernetes. Launch and manage distributed training jobs without hand-rolling the scheduler stack.
DOCSOfficialStructured output and JSON mode with Token FactoryOfficial Token Factory docs showing how to force JSON responses via json_object mode or a strict JSON schema (e.g. a Pydantic BaseModel), with Python, cURL and JavaScript samples.
DOCSOfficialSwitch to Token FactoryMigrate from OpenAI / other inference providers to Token Factory. Drop-in compatibility plus the cost and rate-limit upgrades that come with it.
BLOGOfficialTavily 101: AI-powered search for developersTavily's primer on composing the /search, /extract and /crawl endpoints into real agent workflows, giving LLMs reliable real-time web access for RAG and research.
DOCSOfficialTavily Research API endpoint referenceOfficial Tavily API reference for the Research endpoint, which runs multiple iterative searches, reasons over sources, and returns a detailed research report in a single agentic call.
WORKSHOPOfficialThe Future of Software Development (Nebius x JetBrains)Become an AI-powered engineer through self-paced training in AI-assisted programming. Industry experts and tutors from leading tech companies. Includes 3 months of any JetBrains IDE free.
BLOGOfficialThe invisible architecture behind great chat appsWhat separates a usable chat product from a janky one — caching, routing, streaming, and the production patterns that hide the seams.
BLOGOfficialThe role of compute cluster networking for AI training and inferenceWhy high-speed interconnects (InfiniBand, NVIDIA Quantum-2) make or break large-scale AI workloads, and how Nebius AI Cloud builds fast, reliable GPU cluster networking for distributed training and inference.
DOCSOfficialToken Factory API referenceFull reference for the Token Factory inference API — chat completions, embeddings, fine-tuning, and batch endpoints.
REPOOfficialToken Factory cookbookReference recipes for Token Factory — inference patterns, fine-tuning runs, evals, and production deployment recipes.
DOCSOfficialToken Factory playgroundInteractive playground for trying open models without writing code. Live-edit prompts, swap models, and copy the request as curl/JS/Python.
DOCSOfficialToken Factory quickstartOpenAI-compatible API to start fast. Set up an API key, hit the inference endpoint, and ship your first request in under five minutes.
REPOOfficialTraining & fine-tuning — Serverless JobsRun GPU workloads that produce model artifacts. Recipes for distributed training and fine-tuning via Serverless Jobs.
BLOGOfficialUsing SkyPilot and Kubernetes for multi-node fine-tuning of Llama 3.1Step-by-step guide to running distributed LLM fine-tuning on Nebius AI Cloud GPU infrastructure.
REPOOfficialvLLM examples (Nebius PS services)Reference vLLM deployment examples in the Nebius Professional Services repo. Drop-in configs for production inference workloads.
DOCSOfficialVoice and media — TTS pipelines on ServerlessBuild pipelines that combine batch processing and inference endpoints for voice synthesis and media workflows.
VIDEOOfficialWelcome to Nebius CloudOrientation video for new Nebius Cloud users. Tour of the console and core concepts.
VIDEOOfficialWelcome to Nebius Token Factory2-minute introduction to Nebius Token Factory — the next evolution of AI Studio with OpenAI-compatible API.
BLOGOfficialWhat are GPU clusters and how to choose yours?A practical primer on GPU compute clusters for training, fine-tuning and inference — covering hardware, orchestration (Kubernetes vs Slurm), InfiniBand networking and shared storage, and how to pick the right configuration on Nebius AI Cloud.
VIDEOOfficialWhy AI Coders Like Cursor and Claude Code Fail Long-TermFailure modes of AI coding agents and patterns that hold up in production. Honest critique + remediation.
BLOGOfficialWhy large MoE models break latency budgets — and what speculative decoding changesProduction analysis of how speculative decoding alters latency for large mixture-of-experts inference workloads.
WORKSHOPOfficialWorkshop: Build an Agentic Slack BotDeploy a web-connected AI agent in under 15 minutes. Build a Slack Pricing Assistant that searches competitor pricing with Tavily, runs inference through Nebius Token Factory, and returns structured recommendations. With Colin Lowenberg (Nebius) + Lakshya Agarwal (Tavily).
VIDEOOfficialWorkshop: Running OpenClaw on NebiusLive webinar with Colin Lowenberg + Michal from Nebius demoing OpenClaw end-to-end: install locally with Token Factory, deploy to Nebius Serverless via Docker (with a CLI skill that automates it), configure sub-agents, and Q&A covering cost, model switching, GitHub deployment roadmap, and Tavily for web search.
BLOGAdding Nebius Token Factory to a Rust Agent Without a Custom ProviderTutorial showing how to use Rig's OpenAI-compatible base_url override to wire Nebius Token Factory into a Rust LLM agent — no custom provider code needed.
WORKSHOPAgentic Search with Tavily — AI Build & LearnA 50-minute hands-on live build using Tavily, the search API built for AI agents and LLMs, to retrieve real-time grounded web results, with a demo and practical community examples.
REPOAI in Mathematical ResearchCourse materials for AI in Mathematical Research. Lean-based exercises and lectures on AI-assisted theorem proving and mathematical discovery.
REPOBoltz-2 Inference on Managed KubernetesStep-by-step manifests + commands to run Boltz-2 protein structure inference at scale on Nebius Managed Kubernetes. Pairs with the YouTube tutorial.
BLOGBuild a Job-Finding Agent with Google ADK, Nebius AI, Mistral OCR & LinkupA multi-agent pipeline that reads resume PDFs with Mistral OCR, searches live job boards via Linkup, and uses Qwen3-14B on Nebius AI Studio to generate and filter matches, orchestrated with Google ADK.
VIDEOBuild a RAG Chatbot to Chat with Your Code with LlamaIndex, DeepSeek v3 & NebiusStep-by-step build of a Retrieval-Augmented Generation chatbot that indexes and queries your codebase using LlamaIndex and the DeepSeek v3 model served on Nebius.
BLOGBuilding a Multi-Agent RAG System with Couchbase, CrewAI, and Nebius AI StudioBuild a semantic search engine that pairs Couchbase as the vector store with CrewAI multi-agent RAG, using Nebius AI Studio for both the Llama LLM and the e5-mistral embeddings.
REPOClaude Code → OpenAI API ProxyProxy Claude Code requests through any OpenAI-compatible endpoint — point it at Nebius Token Factory and run Claude Code on open-source models.
REPOCodex CLI + NebiusWire OpenAI's Codex CLI to Nebius Token Factory. Use Codex's coding workflow with open-source models at lower cost.
REPOConTree MCP ServerRun code in isolated cloud containers. ConTree gives AI agents secure sandboxed execution environments with full root access, network, and persistent images. MCP-compatible.
REPOConTree SkillClaude Code skill for using ConTree — sandboxed container execution with Git-like branching of agent state.
REPOEnd-to-end Slurm training + vLLM inference demo on NebiusA community reference project that provisions a 2-node, 16x H100 Soperator (Slurm-on-Kubernetes) cluster with Terraform, runs SFT and LoRA fine-tuning via sbatch, then serves the model with single- and multi-node vLLM — with reported accuracy gains from 2% to 88%.
VIDEOExploring Nebius Token Factory | Open LLMs, AI Agents, Batch Inference & Fine-TuningA tour of Nebius Token Factory covering its OpenAI-compatible API, batch inference, fine-tuning, and how open models like Qwen Coder plug into agent frameworks and tools like Hugging Face and OpenRouter.
BLOGFine-Tune Your LLM in Minutes with NebiusA practical guide to fine-tuning open-source LLMs on Nebius three ways: the no-code Web Console, the Python SDK, and raw cURL API requests, with .jsonl dataset prep.
BLOGHow I Built an Agentic RAG App to Brainstorm Conference Talk IdeasCombine Tavily live web research, Couchbase vector search over past KubeCon talks, and Nebius AI Studio (e5-mistral embeddings + Qwen3) to synthesize unique conference talk abstracts.
VIDEOHow to Fine-Tune GPT-OSS 20B using Nebius Token FactoryHands-on tutorial fine-tuning the GPT-OSS 20B model on Nebius Token Factory: environment setup, JSONL dataset upload, creating a fine-tuning job, monitoring training, and downloading artifacts.
VIDEOHow to Fine-Tune Open Source LLMs with Nebius Token Factory | Full TutorialEnd-to-end walkthrough of fine-tuning an open-source LLM with LoRA on Nebius Token Factory and deploying it as a production-ready API endpoint, including dataset prep and job configuration.
BLOGI Built a Team of 5 Agents Using Google ADK, Meta Llama and Nemotron-Ultra-253BBuild an AI Trend Analyzer with five sequential ADK agents (Exa, Tavily, Firecrawl + summary/analysis) running Meta Llama 3.1 and Nemotron-Ultra-253B served through Nebius AI Studio.
BLOGI Used Agent Skills to Fine-Tune an Open-Source LLM on Nebius Token FactoryA teacher-student distillation walkthrough for an insurance-claims chatbot, using Token Factory Data Lab batch inference, LoRA fine-tuning, serverless adapter deployment, and a Gradio comparison app. Ships with a companion Jupyter notebook.
REPOLangChain + NebiusConfigure Nebius Token Factory as the inference provider in LangChain apps. Drop-in replacement for OpenAI client.
REPOLLM Engineering EssentialsMaterials for the LLM Engineering Essentials course. Notebooks covering training, fine-tuning, evals, and deployment patterns.
REPOLLM Engineering Knowledge BaseSupporting reference materials for the LLM Engineering Essentials course. Lookup-style HTML pages organized by topic.
REPOLLMOps EssentialsLLMOps educational project — covers deployment, monitoring, and lifecycle management for LLM systems in production.
REPOMachine Learning Starter PackOpen ML course — fundamentals through hands-on Jupyter notebooks. Good first stop for anyone newer to ML before going deeper.
REPONebius MCP ServerModel Context Protocol server for Nebius. Lets MCP-aware clients (Claude Code, OpenClaw) operate Nebius resources as native tools.
REPONebius Skill (Claude Code + OpenClaw)Dual-compatible Claude Code + OpenClaw skill for managing Nebius AI Cloud infrastructure via CLI. Drop it in, get instant Nebius CLI mastery in your agent.
REPOOpenClaw + NebiusOpenClaw provider plugin for Nebius, plus the deployment UI (claw.moi) and the Claude Code skill. The reference implementation if you want to wire OpenClaw to Nebius end-to-end.
REPOOpenClaw Nebius Token Factory PluginToken Factory provider plugin for OpenClaw — 44+ open-source models via a single endpoint. The TypeScript provider that powers the rest of the OpenClaw + Nebius stack.
REPOQLoRA Fine-tune Gemma 4 on NebiusValidated recipe + Nebius infra skill + 9-node workshop fleet runbook. Built and run live at Sandbox VR SF (May 2026, ~30 attendees fine-tuned hands-on).
REPOSolutions · Anyscale on NebiusRun Anyscale (managed Ray) on Nebius infrastructure. Terraform module for the integration plus reference workloads.
REPOSolutions · GPU SlicingMulti-tenant GPU slicing setup for Nebius — partition a GPU across workloads with MIG / time-slicing. For shared dev environments.
REPOSolutions · Kubernetes Training ClusterTerraform module for a Nebius Managed Kubernetes cluster optimized for distributed training. GPU node pools, NCCL networking, fast storage.
REPOSolutions · Soperator (Slurm on Kubernetes)Reference Terraform setup for Soperator — Nebius's Slurm-on-Kubernetes operator. Deploy a Slurm cluster on managed K8s with one terraform apply.
REPOSolutions · VM Instance (Terraform)Baseline Terraform module for a Nebius VM. Building block — fork it as the starting point for your own infra modules.
BLOGText-to-SQL: Creating Embeddings with Nebius AI Studio (Part 1)Part 1 of a text-to-SQL RAG series: turn SQL schema into annotated markdown and generate vector embeddings with Nebius AI Studio (BAAI/bge-en-icl), stored in Postgres with pgvector.
BLOGText-to-SQL: Generating SQL with Nebius AI Studio (Part 2)Part 2 of the text-to-SQL RAG series: use the embeddings from Part 1 to retrieve relevant schema and generate correct SQL queries with Nebius AI Studio models.
BLOGText-to-SQL: Querying Databases with Nebius AI Studio and Agents (Part 3)Part 3 of the text-to-SQL RAG series: wrap the pipeline in an agent that queries a live database end to end, powered by Nebius AI Studio models.
REPOTF Cookbook · Building AgentsRecipes for building agents on Nebius Token Factory. Tool-using LLMs, memory patterns, and orchestration examples.
REPOTF Cookbook · DistillationModel distillation recipes — teach smaller, cheaper models from a larger teacher. Run on Nebius compute, serve via Token Factory.
REPOTF Cookbook · LoRA Fine-tuningLoRA / QLoRA fine-tuning recipes for Token Factory models. Includes sample datasets, training scripts, and serving the adapter.
REPOTF Cookbook · Post-TrainingPost-training techniques for open-source models on Nebius — RLHF, DPO, instruction tuning. End-to-end working examples.
REPOTF Cookbook · RAG PatternsRetrieval-augmented generation recipes using Token Factory inference. Vector stores, chunking, hybrid retrieval, and evals.
REPOTF Cookbook · Tool CallingTool-calling patterns on Token Factory. Schema design, parallel calls, error handling, structured outputs.
BLOGUse DeepSeek R1 & V3 with Bolt.DIY & Cursor in 3 StepsGet free Nebius AI Studio API keys, route DeepSeek R1/V3 through OpenRouter, and plug the EU-hosted models into Bolt.DIY and Cursor for coding.
BLOGUsing Nebius AI Models with LangChain/LangGraph via LiteLLMWire Nebius-hosted Qwen models into LangChain/LangGraph through LiteLLM, then build a ReAct agent that talks to databases over the Model Context Protocol.
VIDEOWeb Search MCP Server | Tavily & LangGraph | Full Tutorial With CodeDeploy the Tavily MCP server with NPX, expose it over streamable HTTP via Supergateway, wire up a LangChain MCP client, and connect a LangGraph agent for live web search.
