OpenResearch

What is OpenResearch?

OpenResearch is a tool from the alphaXiv team that turns your favorite coding agents into research agents able to review literature, develop hypotheses and run experiments. A local CLI called orx structures the experiments on your machine, while a compute marketplace lets you compare live GPU and CPU prices and availability across more than a dozen providers and launch an instance in one click. Instances come preinstalled with uv and Claude Code. Code, results and experiment history remain in your local repository, and OpenResearch provisions the compute and synchronizes access. You can use the web dashboard, the macOS app for Apple silicon and Intel, or the open-source CLI, and you can also connect an existing machine over SSH. Hosted compute is prepaid and billed per millisecond at the underlying provider rate with no OpenResearch markup, with featured GPUs such as the RTX 4090 starting around $0.89 an hour.

Use cases

• Run autonomous research experiments with coding agents
• Rent on-demand GPUs for model training
• Compare GPU prices across cloud providers
• Review literature and form hypotheses with an agent
• Manage SSH machines for experiments
• Keep experiment history in a local repository

Who it's for

• ML researchers
• Graduate students
• AI engineers
• Independent researchers

Why use OpenResearch

• Prices are passed through at the provider rate, so you pay no OpenResearch markup on hosted compute.
• Code, results and experiment history remain in your own repository.
• Comparing providers in one place saves checking each cloud separately.
• Preinstalled tooling gets an instance ready for agent-driven experiments quickly.
• You can bring your own machines or provider accounts instead of using hosted compute.

OpenResearch features

• orx CLI for browsing offers, registering SSH keys and managing instances
• Compute marketplace comparing GPU and CPU prices across providers
• One-click launch of hosted instances
• Instances preinstalled with uv and Claude Code
• Startup-time and availability metrics for each provider
• Per-millisecond billing at provider rates with no markup
• Connect existing machines over SSH or use your own provider credentials
• Web dashboard with an Instances tab showing SSH details and lifecycle controls
• macOS app that runs the local dashboard

OpenResearch integrations

• Claude Code
• uv
• SSH

How to get started with OpenResearch

1. Install the macOS app or run the install script for the orx CLI, then sign in with orx login.
2. Register an SSH key and browse live GPU and CPU offers in the Compute tab or CLI.
3. Launch a hosted instance, or connect an existing machine over SSH.
4. Point your coding agent at the project to review literature, test hypotheses and run experiments.
5. Track balance and charges in the Billing tab and stop instances when finished.

OpenResearch limitations

• Hosted compute is prepaid, and a zero balance blocks new instances and ends running ones
• The macOS app requires macOS 11 or later and does not put orx on your PATH
• Self-managed machines are billed by their own provider, not through OpenResearch

OpenResearch pricing

OpenResearch is open source, so you can use and self-host it for free.

OpenResearch plans

• Hosted compute

  Usage-based

  prepaid, billed per millisecond at provider rate

  • No OpenResearch markup
  • RTX 4090 from $0.89/hr
  • H100 SXM $2.73/hr
  • Top up from the Billing tab

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

OpenResearch alternatives

The main OpenResearch alternatives listed on Debuly are Lambda, Runpod, Vast.ai, Modal, Chart Library and Consensus. The table compares pricing and platforms; order is A–Z.

1. 

   Chart Library

   Lookup tool and MCP/API service that finds historical analogs for a stock's recent chart and shows what followed.

   Research and searchDeveloper platformsAutomation and agents
   FreeVisit Chart Library (official site)
2. 

   Consensus

   AI search engine for research that answers questions from peer-reviewed papers, with every claim linked back to its source.

   Research and search
   FreemiumVisit Consensus (official site)
3. 

   Lambda

   GPU cloud for AI with on-demand NVIDIA instances and reserved 1-Click Clusters for training, fine-tuning and inference.

   Developer platformsCoding
   PaidVisit Lambda (official site)
4. 

   Modal

   Serverless cloud platform for running AI inference, training, batch jobs and agent sandboxes, billed per second of compute.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Modal (official site)
5. 

   Runpod

   Pay-per-second GPU cloud for AI: on-demand Pods, Serverless endpoints, multi-node Clusters and ready-made public model APIs.

   Developer platformsCoding
   PaidVisit Runpod (official site)
6. 

   Vast.ai

   GPU cloud marketplace for renting on-demand, interruptible or reserved GPUs with per-second billing.

   Developer platformsAutomation and agents
   PaidVisit Vast.ai (official site)

OpenResearch FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from openresearch.sh · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.