tensorlake.ai

免费大模型额度 · 免费模型 · 社区口碑 · 风险评估

网站类型其他/无关
状态可用
口碑分60 / 100
风险等级
置信度85%
发现来源12britz/awesome-free-models

网站介绍

Tensorlake 是一家位于旧金山的 AI 基础设施初创公司(成立 2022,累计融资约 $8M,种子轮),主打两条产品线:面向 AI agent 的文档摄取/解析数据云(把 PDF、图片、幻灯片等非结构化文档转成结构化 JSON/Markdown,按页计费,约 $0.01/页)以及面向 agent 的「沙箱原生云」——用 Firecracker MicroVM 安全运行 LLM 生成的不可信代码并做编排。它不是免费大模型 API 提供方,本身不提供可调用的大模型推理接口,用户需自带模型(如 Anthropic Claude)接入其基础设施。

免费额度

官网标注「Free tier included,无需信用卡」可免费试用文档摄取/沙箱,但具体免费额度未公开明示;文档解析正式定价约 $0.01/页起。注意:此免费层是文档/沙箱基础设施的免费试用,并非免费大模型推理 API。

免费模型

未知

优点

缺点

风险点

社区口碑综述

Tensorlake 在公开渠道(Crunchbase、PitchBook、Tracxn、StartupHub、G2、SourceForge 等)有正常的公司画像与产品评测,是一家有融资、有官方文档与开源 SDK 的正规初创公司,无跑路/盗 key 等负面信号。但它与「免费大模型 API」这一主题无关,社区中也没有把它当作免费 LLM 中转/供给来讨论。声誉评分仅反映其作为正规公司的可信度,不代表它属于 airouter 收录类别。

使用建议

不收录为免费大模型 API 站点。Tensorlake 是面向 AI agent 的文档摄取/解析数据云与沙箱运行时基础设施,需用户自带大模型,不提供免费可调用的大模型推理 API,应归类为 other。若用户需要的是文档解析/agent 沙箱基础设施可单独评估,但与本项目「免费大模型 API」目标无关。

最近探测

可达是 (HTTP 200)
延迟1406 ms
平台unknown
模型端点需鉴权

社区提及 (2)

Homepage probe (tensorlake.ai) · neutral
Title: 'Tensorlake — Sandbox-native cloud for AI agents.' Homepage describes isolated stateful sandboxes (Firecracker MicroVMs), lifecycle management, orchestration, remote dev and a document ingestion engine. Explicitly infrastructure-focused — 'not an LLM provider'; users bring their own models. 'Free tier included. No credit card required.' No public models, no callable inference API.
Web search / company profiles (Crunchbase, StartupHub, Tracxn) · neutral
Tensorlake — San Francisco, founded 2022, ~$8M raised (seed). 'AI data cloud that reliably transforms data from unstructured sources into ingestion-ready formats for AI applications.' Document parsing billed from ~$0.01/page; described as document ingestion + agent sandbox infrastructure, not a free LLM API.

参考来源

In English

Summary: Tensorlake is a San Francisco AI-infrastructure startup (founded 2022, ~$8M seed funding) with two main product lines: a document ingestion/parsing data cloud for AI agents (turning PDFs, images and slides into structured JSON/Markdown, billed per page at ~$0.01/page) and a 'sandbox-native cloud' for agents that uses Firecracker MicroVMs to securely run untrusted LLM-generated code and orchestrate workflows. It is NOT a free LLM API provider — it offers no callable model-inference endpoint; users bring their own models (e.g. Anthropic Claude) and run them on Tensorlake's infrastructure.

Free quota: The site advertises a 'Free tier included, no credit card required' for trying document ingestion/sandboxes, but the exact free allowance is not publicly specified; document parsing is officially priced from ~$0.01/page. Note: this free tier is a trial of the document/sandbox infrastructure, not a free LLM inference API.