DOCSEARCHbyAlgolia
Submit Docs

Fluence Documentation

website·Indexed 33 days ago

https://fluence.dev

Documentation for the Fluence decentralized compute marketplace, featuring guides for CPU and GPU Cloud instance rental, API integration, and provider server management. Includes details on the Fluence Network, staking with FLT tokens, the Fluence Bridge to Ethereum, and programmatic access via the Fluence API.

Tokens
15.8K
Snippets
30
Records
124
Agent score
69%

What's inside Fluence

  1. Overview of Fluence Bridge and Network Concepts

    Fluence is a Layer 2 (L2) network that uses an Optimistic Rollup system to scale Ethereum. Key concepts include:

    • Optimistic Rollup: Assumes transactions are valid unless proven otherwise via 'fraud proofs'.
    • Challenge Period: A 7-day waiting period for withdrawals from Fluence to Ethereum to allow for the identification and dispute of fraudulent transactions.
    • Transfer Timelines: Deposits (Ethereum to Fluence) take up to 15 minutes. Withdrawals (Fluence to Ethereum) take 7 days due to the challenge period; funds must be manually claimed in the Bridge App after this period ends.

  2. Overview of paying with x402 for agents

    Fluence supports the x402 protocol, allowing HTTP clients to top up their account balance using USDC on the Base network. This process removes the need for manual signup, emails, or console visits, as the paying wallet serves as the account identity. The general workflow for an agent to get compute is: top up via x402, sign in using the wallet (SIWE), call the Fluence API with the resulting token, and rent a VM.

  3. Overview of the Fluence Provider Application

    The Fluence Provider Application is the centralized tool used by Providers on the Fluence Network to manage core operations. It handles three primary areas:

    1. Infrastructure Management: Managing Talos Linux (the supported OS), provisioning Kubernetes (K8s) clusters, and configuring the Custom Fluence Stack for automatic service updates.
    2. Hardware Networking: Configuring private networking (VLANs, Private Subnets) and assigning public IP subnets for customer workloads (e.g., Virtual Machines).
    3. Blockchain Smart-contracts (Expected Q4 2025): Managing Offer contracts (resource registration and pricing), Capacity commitments (staking FLT collateral), and Deal contracts (revenue from infrastructure rentals).

  4. Overview of CPU Cloud VM Resources

    Virtual Machines in CPU Cloud are composed of three separately billed resources located within a specific cluster (data center). All resources are billed per second. Note that terminating a VM does not automatically delete its boot disk or public IP; these must be deleted separately to stop incurring charges.

    • VM Configuration: Defines vCPU and RAM (e.g., cpu-shared-2vcpu-2gb).
    • Boot Disk: A storage volume created from an OS image.
    • Public IP: An IPv4 address for internet access.

  5. Overview of GPU Cloud API

    The GPU Cloud API allows developers to browse GPU plans, deploy instances, and manage their lifecycle. It supports three workload types: containers, VMs, and bare metal. All requests are made against the base URL: https://api.fluence.dev/gpu.

  6. Overview of the Fluence Staking Application

    The Fluence Staking Application is a web interface located at stake.fluence.network that allows users to browse Capacity Commitments (CCs), stake FLT tokens, and monitor performance and rewards.

    Key interface components include:

    • Navigation Bar: Links to the blockchain Bridge Application.
    • Wallet Connection: Button to connect a wallet (required for staking and withdrawing rewards).
    • Capacity Commitments Panel: Shows the distribution of activated CCs by status.
    • Rewards Panel: Summary of earned rewards.
    • Available for Staking: A list of CCs open for delegation.
    • Staked Capacity Commitments / Your Capacity Commitments: Real-time performance data for active CCs. When logged in, this section allows users to withdraw rewards and collateral.

  7. Manage blockchain operations during Provider Application transition

    The initial release of the Provider Application does not support direct interaction with Fluence's core smart contracts. Until this functionality is integrated, providers should continue using the existing Fluence CLI (FCLI) tooling to perform blockchain operations, such as creating Compute Clusters (CC) and withdrawing rewards.

  8. Manage SSH keys using the Fluence API

    SSH keys are account-level credentials used during the deployment of compute instances. All requests to the SSH key endpoints require an access token in the Authorization: Bearer header. The base URL for these requests is https://api.fluence.dev.
    GET /v1/ssh_keys
POST /v1/ssh_keys
DELETE /v1/ssh_keys/{ssh_key_id}
POST /v1/ssh_keys/delete

  9. Manage Fluence account prepaid balance

    Fluence accounts use a prepaid balance in USD to pay for CPU and GPU instance rent. This balance is used exclusively for Fluence resources and cannot be withdrawn. If a refund is required, users must contact the Fluence team.

  10. Manage Fluence resources programmatically via API

    The Fluence API allows for the automation of compute resource management. It can be used to search for available resources on CPU and GPU Cloud marketplaces, deploy virtual machines, GPU containers, and bare metal instances, manage active deployments, and handle SSH keys for instance access.

  11. Understand Fluence CPU Cloud marketplace architecture

    Fluence CPU Cloud is a decentralized compute marketplace for renting VM instances from enterprise-grade providers. The marketplace is structured into the following layers:

    • Offer: A provider listing containing one or more compute peers with identical specifications and pricing in the same data center.
    • Compute peer: The actual physical server that runs the workload.
    • Resource: Specific components on a peer, including vCPU, RAM, storage, or public IP.
    • Hardware specification: Metadata describing the resource (e.g., CPU architecture, memory generation).

    Resources are allocated in multiples of a compute unit (2 vCPUs and 4 GB RAM), with a minimum storage requirement of 25 GB.

  12. Choose between Fluence CPU Cloud and GPU Cloud APIs

    The Fluence API is split into two primary cloud platforms based on the workload requirements:

    • CPU Cloud: Used to search the decentralized CPU compute marketplace and manage virtual machines.
    • GPU Cloud: Used to browse GPU plans and manage containers, VMs, and bare metal instances.

    Account-level management, such as registering and managing SSH keys for deployment, is shared across both platforms.