Doubleword

What it does

Doubleword provides hosted inference for open-weight models through an OpenAI-compatible API. Developers can send model prompts and requests to the service and receive inference outputs; documented options include real-time endpoints for interactive testing, asynchronous inference and 24-hour batch processing. The homepage also lists prompt caching, tool calling and structured outputs. It positions these capabilities for production applications and agent workloads, as well as batch jobs such as evaluation, dataset generation and data annotation.

Self-service access uses an API key and prepaid usage credits. Dedicated inference is separately arranged for workloads needing reserved capacity, custom models or engineering support. Choose realtime, asynchronous or 24-hour batch execution according to the task’s latency needs. Prices vary by model and service speed; consult the live rate schedule before submitting a large job. Region-locked processing is an option for dedicated deployments, rather than a stated default for every request.

What it can do

• Hosted inference for open-weight models
• Real-time inference endpoints
• Asynchronous inference
• 24-hour batch inference
• Prompt caching
• Tool calling and structured outputs

Works with

• OpenAI-compatible API

Before you choose

• Self-service inference is usage-based with prepaid credits; exact credit purchase terms and model rates are not shown in the rate schedule.
• Dedicated capacity is a separate contact-led offering, not the same access route as self-service.
• The homepage describes zero data retention as included in self-service, but vendor documentation does not include its terms or scope.

Listed under

• Coding
• MCP and agent tools
• Async inference
• Batch inference
• Cost reduction
• Developers
• Inference API
• LLM API

Get started

The homepage directs users to get an API key for the inference APIs.

1. Get an API key through the Doubleword app.
2. Choose an available open-weight model and inference mode.
3. Send requests through the OpenAI-compatible API; consult model and pricing documentation for exact rates before use.

Pay-as-you-go with prepaid credits; model, input/output token volume and inference speed affect cost. Dedicated capacity and volume terms are quoted separately.

Source: doubleword.ai, checked 1 Nov 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 1 Nov 2026. Each source supports part of the overview; availability and prices can change.

• Doubleword | Open-Weight Inference for Agents at Scale ↗doubleword.ai

Company website

doubleword.ai: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

manual_review

Last checked 1 Nov 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.