1 / 3
【Overview】 Chutes is a decentralized, serverless inference platform for open-source AI models. Developers can call hosted models through an OpenAI-compatible chat completions endpoint at llm.chutes.ai, or deploy their own containerized models on confidential GPUs. It is aimed at developers and teams who want production access to open-source models without managing the infrastructure behind them. Security sits at the center of the platform. Models can run inside Intel TDX confidential VMs with NVIDIA Protected PCIe GPUs, are attested before execution, and request content is not logged. Only models reporting confidential_compute: true run inside those trusted execution environments. The platform also covers more than text: the catalog includes image, video, speech, music and embedding models, plus moderation models such as Halo guard for hate speech and NSFW detection. 【Key features】 • OpenAI-compatible inference: Serves state-of-the-art open-source LLMs through a chat completions endpoint at llm.chutes.ai. • TEE/secure compute: Runs models in Intel TDX confidential VMs on NVIDIA Protected PCIe GPUs, attested before execution, with no request content logging. • Private Chutes deployment: Runs your own containerized model on dedicated confidential GPU capacity with your own API endpoints. • Multi-modal model catalog: Supports image, video, speech, music and embedding models alongside LLMs. • Batch and ephemeral jobs: Handles ephemeral jobs, batch processing jobs and custom training workloads. • Content moderation models: Includes moderation models such as Halo guard for hate speech and NSFW detection. 【Use cases】 1. LLM chat completions 2. Generate images, video, speech and music 3. Embedding data for search and recommendation 【Pricing】 Plus: $10/month — 2,000 requests/day included, 6% off PAYG rates beyond the quota, $50 monthly usage cap, $4.17 4-hour burst limit Pro: $20/month — 5,000 requests/day included, 10% off PAYG rates beyond the quota, $100 monthly usage cap, $8.33 4-hour burst limit Enterprise: Custom billing only — volume discounts, custom rate limits, dedicated support 【Pros and cons】 Pros: - The OpenAI-compatible endpoint lets teams switch from other APIs with minimal code changes, which keeps migration and tooling costs low. - Confidential compute with pre-execution attestation and no request-content logging suits teams handling sensitive prompts. - A single platform spanning chat, image, video, speech, music, embeddings and moderation reduces the number of vendors a product team has to manage. Cons: - Confidential execution is limited to models that report confidential_compute: true. - Private Chutes deployment is billed per second at the GPU hourly rate plus a one-time deployment fee of 3× the hourly rate. - Monthly subscription plans are not suited to automated or high-volume workloads. 【FAQ】 Q: How is Chutes billed? A: Pay-as-you-go is the default: you top up your account balance and pay only for what you use. Monthly plans (Plus at $10/month, Pro at $20/month) add daily request quotas and discounts, and Enterprise is custom billing only. Q: Who is Chutes built for? A: AI developers, teams shipping AI-powered products in production, and organizations that need private or confidential inference. Q: Do all models run in confidential environments? A: No. Only models reporting confidential_compute: true run inside Intel TDX confidential VMs with NVIDIA Protected PCIe GPUs.
Plus · $10/month (2,000 requests/day included, 6% off PAYG rates beyond the quota, $50 monthly usage cap) Pro · $20/month (5,000 requests/day included, 10% off PAYG rates beyond the quota, $100 monthly usage cap) Enterprise (Custom billing only, Volume discounts, Custom rate limits)
WPS AI is an AI assistant built into WPS Office that helps users write documents, generate presentation outlines and…
Baidu Comate is an AI coding assistant built by Baidu on its ERNIE model. It goes beyond simple autocomplete: it…
UniConverter is an AI media conversion and enhancement tool for video, audio, and images. It combines everyday…
6 similar tools to consider instead of Chutes