openapi: 3.2.0 info: title: Walton MPP Capacity API version: 4.0.0 description: MPP marketplace for GPU capacity, hosted model inference, edge compute, shared workspaces, and supervised resources. Walton coordinates payment and direct handoff metadata without relaying provider capacity traffic. servers: - url: https://api.walton.bot tags: - name: Capacity paths: /capacity/assets: post: summary: Create a provider listing security: - bearerAuth: [] requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/CapacityListingRequest' responses: '201': description: Created listing tags: - Capacity operationId: postCapacityAssets x-operation-id-source: derived /capacity/sessions/{sessionId}/peering-offer: post: summary: Publish provider peering details description: Provider publishes session-scoped public endpoint information and setup steps after reviewing the paid buyer request. security: - bearerAuth: [] parameters: - name: sessionId in: path required: true schema: type: string format: uuid requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/ProviderPeeringOffer' responses: '200': description: Provider handoff ready content: application/json: schema: $ref: '#/components/schemas/CapacitySession' tags: - Capacity operationId: postCapacitySessionsBySessionIdPeeringOffer x-operation-id-source: derived components: schemas: CapacitySession: type: object properties: id: type: string format: uuid status: type: string statusLabel: type: string asset: $ref: '#/components/schemas/CapacityOffer' task: type: string amountCents: type: integer formattedAmount: type: string paymentRail: type: string paymentReceipt: type: string accessToken: type: - string - 'null' handoffUrl: type: - string - 'null' resourceAccess: type: - object - 'null' additionalProperties: true gpuAccess: type: - object - 'null' additionalProperties: true events: type: array items: type: object additionalProperties: true cancellable: type: boolean peering: oneOf: - $ref: '#/components/schemas/PeeringExchange' - type: 'null' CapacityOffer: type: object required: - id - type - title - status - locationLabel - currencyCode properties: id: type: string format: uuid type: type: string enum: - GPU_CAPACITY - MODEL_INFERENCE - EDGE_COMPUTE - SHARED_WORKSPACE - SERVICE_ROBOT - THREE_D_PRINTER title: type: string summary: type: string status: type: string latitude: type: number format: double longitude: type: number format: double locationLabel: type: string hourlyRateCents: type: - integer - 'null' perTaskRateCents: type: - integer - 'null' unitRateCents: type: - integer - 'null' billingUnit: type: - string - 'null' description: Usage unit such as 1K_TOKENS, REQUEST, IMAGE, AUDIO_MINUTE, or KWH. sessionFeeCents: type: - integer - 'null' currencyCode: type: string capabilities: type: array items: type: string gpuModel: type: - string - 'null' gpuCount: type: - integer - 'null' gpuMemoryGb: type: - integer - 'null' gpuAllocationMode: type: - string - 'null' cpuModel: type: - string - 'null' cpuCoreCount: type: - integer - 'null' memoryGb: type: - integer - 'null' storageGb: type: - integer - 'null' connectorType: type: - string - 'null' description: Inference task for MODEL_INFERENCE, workspace type for SHARED_WORKSPACE. connectorProtocol: type: - string - 'null' description: Hosted model endpoint standard or provider-side control protocol. connectorCount: type: - integer - 'null' accessMode: type: - string - 'null' availabilitySchedule: type: array items: type: string requiresOwnerApproval: type: boolean connectorOnline: type: boolean ownerDisplayName: type: string ownerRating: type: number completedRentals: type: integer model: type: - string - 'null' description: Hosted model name or provider model ID for MODEL_INFERENCE. capacity: type: - integer - 'null' description: Concurrent requests for MODEL_INFERENCE or bookable seats for SHARED_WORKSPACE. ProviderPeeringOffer: type: object required: - connectionMode properties: connectionMode: type: string enum: - WIREGUARD - SSH - TAILSCALE - HTTPS_API - MANUAL endpointHost: type: - string - 'null' endpointPort: type: - integer - 'null' minimum: 1 maximum: 65535 providerPublicKey: type: - string - 'null' providerAllowedIps: type: - string - 'null' sshUsername: type: - string - 'null' serviceUrl: type: - string - 'null' format: uri accessCode: type: - string - 'null' providerInstructions: type: - string - 'null' configurationSnippet: type: - string - 'null' providerCompletedSteps: type: array items: type: string PeeringExchange: type: object properties: status: type: string connectionMode: type: string providerReady: type: boolean buyerConfirmed: type: boolean buyerPublicKey: type: - string - 'null' buyerSshPublicKey: type: - string - 'null' buyerEndpoint: type: - string - 'null' buyerAllowedIps: type: - string - 'null' buyerIdentity: type: - string - 'null' buyerNetworkNotes: type: - string - 'null' providerEndpointHost: type: - string - 'null' providerEndpointPort: type: - integer - 'null' providerPublicKey: type: - string - 'null' providerAllowedIps: type: - string - 'null' sshUsername: type: - string - 'null' serviceUrl: type: - string - 'null' accessCode: type: - string - 'null' providerInstructions: type: - string - 'null' configurationSnippet: type: - string - 'null' providerSetupSteps: type: array items: type: string buyerSetupSteps: type: array items: type: string providerCompletedSteps: type: array items: type: string CapacityListingRequest: type: object required: - type - title - summary - autonomyLevel - latitude - longitude - locationLabel properties: id: type: string format: uuid type: type: string enum: - GPU_CAPACITY - MODEL_INFERENCE - EDGE_COMPUTE - SHARED_WORKSPACE - SERVICE_ROBOT - THREE_D_PRINTER title: type: string summary: type: string status: type: string latitude: type: number format: double longitude: type: number format: double locationLabel: type: string hourlyRateCents: type: - integer - 'null' perTaskRateCents: type: - integer - 'null' unitRateCents: type: - integer - 'null' billingUnit: type: - string - 'null' description: Usage unit such as 1K_TOKENS, REQUEST, IMAGE, AUDIO_MINUTE, or KWH. sessionFeeCents: type: - integer - 'null' currencyCode: type: string capabilities: type: array items: type: string gpuModel: type: - string - 'null' gpuCount: type: - integer - 'null' gpuMemoryGb: type: - integer - 'null' gpuAllocationMode: type: - string - 'null' cpuModel: type: - string - 'null' cpuCoreCount: type: - integer - 'null' memoryGb: type: - integer - 'null' storageGb: type: - integer - 'null' connectorType: type: - string - 'null' description: Inference task for MODEL_INFERENCE, workspace type for SHARED_WORKSPACE. connectorProtocol: type: - string - 'null' description: Hosted model endpoint standard or provider-side control protocol. connectorCount: type: - integer - 'null' accessMode: type: - string - 'null' enum: - WIREGUARD - SSH - TAILSCALE - HTTPS_API - MANUAL - null description: Default direct handoff mode offered by the provider. availabilitySchedule: type: array items: type: string requiresOwnerApproval: type: boolean connectorOnline: type: boolean deprecated: true description: Legacy field retained for compatibility. ownerDisplayName: type: string ownerRating: type: number completedRentals: type: integer exactLocationLabel: type: - string - 'null' description: Private provider location. Never returned by public offer search. accessInstructions: type: - string - 'null' description: Provider-side preparation notes. Public searches do not expose private setup secrets. apiControlEnabled: type: boolean deprecated: true description: Legacy connector flag. Direct peering does not require a Walton connector. mppEnabled: type: boolean depositCents: type: - integer - 'null' maxConcurrentSessions: type: - integer - 'null' containerRuntime: type: - string - 'null' model: type: - string - 'null' description: Hosted model name or provider model ID for MODEL_INFERENCE. capacity: type: - integer - 'null' description: Concurrent requests for MODEL_INFERENCE or bookable seats for SHARED_WORKSPACE. securitySchemes: bearerAuth: type: http scheme: bearer x-walton-model-providers: - Z.ai - Moonshot AI (Kimi) - DeepSeek - Mistral AI - Cohere - Anthropic - OpenAI - Google - Meta - Alibaba Cloud (Qwen) - xAI - MiniMax - AI21 Labs - Jina AI - Nomic AI - Nous Research - Black Forest Labs - Stability AI - Microsoft - Other