🌐 English · 中文

api.ranmeng.icu

Free LLM quota · free models · community reputation · risk review

TypeFree public relay
StatusUp
Platformopenai-compatible
Reputation42 / 100
Risk level Med
Confidence40%
Websitehttps://ranmeng.icu ↗
Discovered fromrouterpark

Overview

Jianghu Yizhan (api.ranmeng.icu, part of the 'Ranmeng Tech' resource site ranmeng.icu) describes itself as 'a non-profit API relay for AI learners and individual developers, offering small learning/test quotas' — free, not for sale, with no commercial SLA promised. It is anonymously operated, reachable only via a QQ group, and payments are currently disabled.

Free quota

Non-profit and free, offering small learning/test quotas (the site states 'resources are limited, take only what you need; not for sale, not for resale, no commercial SLA promised'). Registration requires email verification and only @qq.com addresses; payment is disabled (payment_enabled=false). Exact daily/total quota is undisclosed; /v1/models requires auth (401), so details are unknown.

Free models

Unknown

Pros

Cons

Risk flags

Community reputation

Almost no independent community reputation: V2EX, Linux.do, NodeSeek, Zhihu, and GitHub return no usable discussion of Jianghu Yizhan / api.ranmeng.icu; it can only be judged from its own front-end __APP_CONFIG__ and terms. Compared with similar anonymous sites it is more honest in positioning (explicitly non-profit, small test quota, no SLA promise) and has a QQ group and fairly complete terms, so trustworthiness is slightly higher; but the operator is still anonymous, reputation is zero, and free top-tier models carry the common stolen-key suspicion — assessed cautiously as an anonymous non-profit new site.

Recommendation

As an honestly self-described anonymous non-profit small site, it can be used for small-scale learning/testing after fully understanding the risks, but not for production or sensitive data, and do not top up or rely on it long-term because of 'free Claude/GPT'. Keep alternative channels ready for its rate-limiting/shutdown, and note that its free top-tier models likely come from an illegitimate source, so use itself may violate upstream terms.

Latest probe

ReachableYes (HTTP 204)
Latency1673 ms
Platformunknown
Models endpointauth required

Mentions (3)

站点前端配置 window.__APP_CONFIG__ · neutral
site_name=江湖驿站；site_subtitle=「面向 AI 学习者和个人开发者的公益 API 驿站，提供小额学习测试额度。资源有限，请按需取用，不售卖、不转卖、不承诺商用 SLA。」；contact_info=QQ群 297354473；registration_enabled=true、email_verify_enabled=true、邮箱白名单=@qq.com；payment_enabled=false；doc_url=ranmeng.icu/api-zhongzhuan。 ↗
服务条款草案（站点 login_agreement） · neutral
「本服务是 AI API 网关和额度分发平台。平台本身不训练基础模型……您的请求可能会被转发至一个或多个上游模型服务商（OpenAI、Anthropic、Google、AWS、Azure 等）」，明示为转发型中转。 ↗
社区检索结果 · negative
V2EX、Linux.do、NodeSeek、知乎、GitHub 均未检索到 江湖驿站 / api.ranmeng.icu 的独立讨论；运营主体匿名，仅 QQ 群联系。

Sources

Community discussion · reputation check

Want to verify api.ranmeng.icu's real reputation before deciding? Read first-hand community discussion:

External discussion quality varies; judge together with the risk review on this page.

FAQ

Is api.ranmeng.icu safe and trustworthy?

Overall reputation 42/100, risk level Med. Almost no independent community reputation: V2EX, Linux.do, NodeSeek, Zhihu, and GitHub return no usable discussion of Jianghu Yizhan / api.ranmeng.icu; it can only be judged from its own front-end __APP_CONFIG__ and terms. Compared with similar anonymous sites it is more honest in positioning (explicitly non-profit, small test quota, no SLA promise) and has a QQ group and fairly complete terms, so trustworthiness is slightly higher; but the operator is still anonymous, reputation is zero, and free top-tier models carry the common stolen-key suspicio…

Will api.ranmeng.icu shut down? Is the service stable?

Known risks: Anonymous operator, reachable only via a QQ group (297354473), with no registration/company entity info; Non-profit site explicitly promises no SLA and has limited resources; may rate-limit, cut quota, or shut down at any time; Offers closed-source large models like Claude/GPT/Gemini for free, suggesting stolen/shared upstream keys; compliance doubtful; No independent community reputation anywhere; trustworthiness cannot be externally verified; Free quota, model list, and rate-limit rules opaque (/v1/models requires auth; the doc site once returned 520); Conversation data relayed…

Which free models does api.ranmeng.icu offer? How much free quota?

Free quota: Non-profit and free, offering small learning/test quotas (the site states 'resources are limited, take only what you need; not for sale, not for resale, no commercial SLA promised'). Registration requires email verification and only @qq.com addresses; payment is disabled (payment_enabled=false). Exact daily/total quota is undisclosed; /v1/models requires auth (401), so details are unknown.. Free models: unknown.

Is api.ranmeng.icu worth using? How should I use it?

As an honestly self-described anonymous non-profit small site, it can be used for small-scale learning/testing after fully understanding the risks, but not for production or sensitive data, and do not top up or rely on it long-term because of 'free Claude/GPT'. Keep alternative channels ready for its rate-limiting/shutdown, and note that its free top-tier models likely come from an illegitimate source, so use itself may violate upstream terms.

Related free-LLM sites