跳到正文

FreeLLMAPI 版本记录

一个本地 OpenAI 兼容路由器,用一个密钥就能使用多家提供方的免费 LLM 模型,支持智能路由、自动故障转移和密钥加密。

Cask开发工具自带更新 5.6万
主页
安装命令
brew install --cask freellmapi
近 30 天安装
94
App榜第 971
近一年安装
94
365 天累计
最新版本
0.16.38
下载大小
127MB
dmg
更新节奏
每天
近 30 天 28 个版本
  1. 0.16.38

    最新补丁

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.38 增加固定模型 ID 的原生 Puter 聊天,以及生图、WAV 转写、翻译和抽取。

    Puter
    原生 Puter 聊天固定模型 ID。拒绝后不再回退到其他模型。
    新能力
    新增 AI Horde 和 xKiro 生图。新增 Google Live WAV 转写。翻译和抽取不是普通聊天路由。

    FreeLLMAPI 0.16.38 于 2026 年 11 月 21 日发布。给把请求转到免费聊天模型的人增加原生 Puter 聊天:固定模型 ID,被拒绝后不改走其他模型,并使用单独的免费 Puter 账号令牌。

    来源:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.38

    • 原生 Puter 聊天固定模型 ID,被拒绝后不回退。使用单独的免费 Puter 账号令牌。
    • 新增 AI Horde 和 xKiro 图像生成。
    • 新增 Google Live WAV 转写。
    • 新增明确的 xKiro 与 Cloudflare 翻译,以及 Clairic Flash 抽取。翻译和抽取不是普通聊天路由。
    • 已有的签名目录访问层级和 30 天推广不变。
  2. 0.16.37

    补丁

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.37 面向用一个兼容 OpenAI 的应用转发本地和托管模型的人,加入 cheapllm.shop,并在服务只返回错误正文时改走下一家。

    服务
    cheapllm.shop 是一家服务:一个免费模型,DeepSeek V4.1 Flash 1017。该模型支持工具、JSON schema 和图像。在密钥页加入密钥。这一行现在是 Premium,30 天后为 Free,经由托管目录。
    路由
    HTTP 200 但正文只是错误时,会按正确的冷却时间改走下一家。这种响应不再算作空回答。

    FreeLLMAPI 0.16.37 面向用一个兼容 OpenAI 的应用转发本地和托管模型的人,加入 cheapllm.shop,并在服务只返回错误正文时改走下一家。

  3. 0.16.36

    补丁

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.36 新增翻译和摘要接口,不进入普通对话,也不参与自动回退。

    接口
    GET /v1/text/models 列出翻译和摘要模型。POST /v1/text/translate 翻译文本。POST /v1/text/summarize 摘要文本。这些路由不进对话,也不自动回退。
    目录
    Premium 和 30 天免费的排期不变。

    FreeLLMAPI 0.16.36 于 2026 年 11 月 19 日发布,新增专门的翻译和摘要接口,不进入普通对话,也不参与自动回退。来源:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.36

    • GET /v1/text/models 列出可以翻译或摘要的模型。
    • POST /v1/text/translate 翻译文本。
    • POST /v1/text/summarize 摘要文本。
    • 这些路由不进入普通对话,也不参与自动回退。
    • Premium 和 30 天免费目录的排期不变。见 https://github.com/tashfeenahmed/freellmapi/pull/2009
  4. 0.16.34

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.34 列出全部 21 个桌面代理,并能连接网关。

    代理
    在 macOS 或 Windows 上打开 Agents,可看到全部 21 个代理。按名称搜索,清空搜索后会重新显示完整列表。可在 Claude Code、Cline、OpenCode 或其他代理旁连接 FreeLLMAPI。应用会保存支持的配置,需要手动步骤时会给出指引。
    修复
    转录上传的音频缓冲已正确处理。

    FreeLLMAPI 0.16.34 于 2026 年 11 月 17 日发布,是给桌面编程代理接本地网关的人用的 OpenAI 兼容路由。这一版列出全部 21 个代理,并能从应用里开始连接。

    来源:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.34

    • 在 macOS 或 Windows 上打开 Agents,可看到全部 21 个代理。
    • 按名称搜索,清空搜索后会重新显示完整列表。
    • 可在 Claude Code、Cline、OpenCode 或其他代理旁连接 FreeLLMAPI。
    • 应用会保存支持的配置,需要手动步骤时会给出指引。
    • 转录上传的音频缓冲已正确处理。
  5. 0.16.33

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    Meldport 对话与 Solaria-3 转写

    FreeLLMAPI 0.16.33 为本地路由免费模型的人增加 Meldport 对 GPT-6 Luna 免费别名的对话,以及无密钥的 OVHcloud Solaria-3 转写。

    FreeLLMAPI 0.16.33 于 2026-11-17 发布,为通过本地 OpenAI 兼容密钥路由免费模型的人增加两条新路线。

    • Meldport 对话支持 GPT-6 Luna 的免费别名。适配器只请求促销额度,核对响应的额度和模型身份,并支持流式输出。Meldport GLM 仍关闭。
    • 无密钥的 OVHcloud Solaria-3 转写接受英文 PCM16 WAV 上传。现有 OVH Whisper 转写仍用原来的端点。
    • 模型可用性仍按已签名目录:Premium 在发布时可用,Free 在 30 天后可用。现有模型路由和目录资格不变。

    来源:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.33

  6. 0.16.32

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.32 于 2026-11-17 发布,为用一把本地密钥路由免费模型的人增加了 SpeechFlow 文件转写和更多嵌入尺寸。来源:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.32

    新增
    SpeechFlow 文件转写需要密钥 ID 和密钥。支持 Free.ai 嵌入,以及另外两种 SilvR Nemotron 尺寸。
    说明
    五条路线走签名目录,本地停用仍然有效。Premium 立即获得,Free 等待 30 天。
  7. 0.16.31

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.31 于 2026-11-16 发布,为用一把本地密钥路由免费模型的人增加了经过测试的聊天、图像和转写路线。

    新增
    NovAI GLM 4.7 Flash 支持普通聊天和流式聊天。ElectronHub FLUX.2 Klein Base 4B 可以生成图像。SilvR Whisper Large V3 可以转写音频。
    说明
    模型身份校验和有界图像解码仍然保留。Premium 立即获得 11 月 16 日目录,Free 需等待 30 天。
  8. 0.16.30

    展开

    GitHub Release Homebrew 收录

    兼容性macOS 12 起

    Native routing for the November 15 renewable-free catalog additions:

    • SilvR AI: three embedding models, Whisper Large V3 Turbo transcription, and Gemini Decision 1 predicate questions.
    • Cartesia Ink 2 transcription: PCM16 WAV upload over the native, proxy-aware WebSocket with finalization checks and no ffmpeg dependency.
    • ElectronHub images: Qwen Image 2512 plus dispatch for the two existing catalog image routes; bounded image decoding and safe timeout/error handling.

    Also includes the current main-branch fixes for Windows bootstrap root resolution, explicit non-streaming chat requests, key health evidence, the Codex model catalog endpoint, and advertising the router's effective context window in model listings. Concurrently merged README contributor credits are preserved.

    Provider free allowances remain shared, renewable and subject to availability. No new app quotas, paid fallback, Mistral catalog changes or premium/Free release-timing changes. Catalog publication is separate: seven additions and one retired Kenari route disabled, with premium access immediately and Free eligibility after the existing 30-day window.

  9. 0.16.29

    展开

    GitHub Release Homebrew 收录

    兼容性macOS 12 起

    New providers

    • Typecast: native SSFM 3.1, 3.0 and 2.1 text-to-speech.
    • Cartesia: Sonic 3.6, 3.5 and 3 speech, plus Ink Whisper batch transcription.
    • SilvR AI: OpenAI-compatible chat with normal and streaming model-identity checks.

    The new providers appear in local key setup/import. Typecast Free output is non-commercial and requires attribution; Cartesia's commercial-use license is on a paid plan. Free allowances are shared account pools, not credits per model.

    Native speech adapters preserve error/backoff information for failover, refuse credential redirects, bound downloads and validate actual audio / transcription responses. Chat cannot accidentally route to a speech model. This release also includes the recent URL-guard, Hugging Face video-download and request-lease fixes already merged into main.

    Model rows are delivered separately through the signed catalog. Premium receives updates immediately; Free receives them after the existing 30-day aging rule. Mistral and local disabled choices are unchanged.

  10. 0.16.28

    展开

    GitHub Release Homebrew 收录

    兼容性macOS 12 起

    v0.16.28 — November catalog adapters

    New catalog adapter support

    • Cloudflare EmbeddingGemma 2: native text embeddings with fixed 768-dimensional vectors, batch ordering and model-identity validation. Uses its own embedding family so existing indexes never silently switch vector spaces.
    • Google Gemini Decisions: native typed decisions using existing Google credentials, with validated answers, confidence and token usage.
    • ElectronHub transcription: Voxtral Mini Transcribe 2 and Deepgram Nova 3 through the existing key pool and multipart speech endpoint.

    Quota errors and retry hints remain available to the normal router. Existing providers, Mistral behavior and user overrides are unchanged.

    The accompanying catalog update is delivered separately: ten additions, three retired routes disabled (net enabled +7), and corrected Speechify allowance metadata. Premium receives the Live catalog immediately; Free receives new entries through the existing 30-day release process.

    Validation

    Production builds, lint, migration round-trip tests, full server/CLI/client suites and desktop tests passed locally. Native adapters have regression coverage for malformed responses, model substitutions, vector dimensions and rate limits. Live catalog tests use the actual local application API without saving provider credentials.

  11. 0.16.27

    展开

    GitHub Release Homebrew 收录

    v0.16.27 — Zro daily-free provider

    Zro by MoonMath.ai

    • Add your own Zro API key in the Keys dashboard; imports recognize ZRO_API_KEY and MOONMATH_API_KEY.
    • OpenAI-compatible normal and streaming chat, tool calls and JSON output use the existing routing and failover behavior.
    • Credential validation uses the authenticated model-list endpoint without spending inference tokens.
    • Reject mismatched response model identities, credential redirects and in-band streaming errors.
    • Account-wide Free usage resets daily; it is not a monthly cash grant. Model availability is managed through the signed catalog rather than bundled seeds.

    Catalog compatibility

    This release supplies the Zro adapter needed by the November 12 catalog update. The existing Premium-immediate / Free-after-30-days catalog logic is unchanged. Mistral and other providers' application behavior are unchanged.

    The catalog publication is separate from the app release. The validated update holds three dated/non-renewing promotions rather than mislabeling them as recurring free, retains stored history, and adds only tested routes.

    Local validation: full server/CLI/client suites, locale validation, migration round trips, lint, workspace build and desktop tests. Zro was also tested through the real app router against the Free account with an in-memory database.

  12. 0.16.26

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.26 修复服务商路由、本地仪表板和模型发现。

    FreeLLMAPI 0.16.26 于 2026-11-11 发布,面向把一个本地接口路由到多家免费和付费语言模型服务的人。这一版修复服务商路由、本地仪表板和模型发现。官方说明:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.26

    • 修复已明确指定服务商时的默认模型路由。
    • 修复端口已被占用时的本地仪表板。
    • 修复 NVIDIA NIM 模型发现和一处缺失类型。
    • 修复 LongCat 图像模型标识和服务商设置面板。
  13. 0.16.25

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeTTS:本地个人语音

    FreeLLMAPI 0.16.25 增加本地 FreeTTS 语音:自备密钥,即可获得可选声音的原生 MP3。

    FreeLLMAPI 0.16.25 于 2026 年 11 月 11 日发布。自己运行 FreeLLMAPI 网关的人可以在本地接入自备密钥的语音:添加个人 FreeTTS 密钥后,即可使用原生 MP3 文本转语音并选择声音。

    来源:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.25

    • 为本地账户增加 FreeTTS 密钥设置与导入。
    • 以 MP3 朗读,支持选择声音、固定主机下载、有界响应、音频校验和限流切换。
    • 目录新增一项 neural-tts。声音是参数,不是单独的模型。
    • 保留自 0.16.24 以来已并入主线的 Gemini 模式兼容、过载模型的备用密钥切换,以及开机启动开关。

    FreeTTS Free 仍限于个人、非商业用途。不附带任何服务商密钥。Premium 现在获得 neural-tts;Free 仍按现有 30 天老化规则之后才获得。

  14. 0.16.24

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.24 为在本地转发 OpenAI 兼容请求的人加入 PowerTokens 作为后备供应商,并让自定义端点的多个 Key 轮换处理图像、语音、转写和嵌入。

    供应商
    PowerTokens 作为 OpenAI 兼容后备加入,glm-4.7-flash 为 $0。遇到 429 会冷却该 Key,路由改走其他来源。
    修复
    自定义端点会轮换图像、语音、转写和嵌入的 Key(#1738)。立即探测为推理模型的隐藏思考留出空间(#1862)。

    FreeLLMAPI 0.16.24 于 2026 年 11 月 10 日发布,为在本地转发 OpenAI 兼容请求的人加入 PowerTokens,并修复自定义端点的 Key 轮换。来源:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.24

    • PowerTokens(api.powertokens.ai)是 OpenAI 兼容供应商;glm-4.7-flash 标价 $0,对话、流式和工具调用已测试。
    • PowerTokens 常返回 429,路由会冷却该 Key 并改走其他来源;请把它当作后备,并在受限 Key 上允许 glm-4.7-flash。
    • 模型经签名目录下发:Premium 立即获得,Free 仍等待现有的 30 天。
    • 自定义端点的多个 Key 在某个 Key 被禁用、超预算、限速或拒绝时,会轮换处理图像、语音、转写和嵌入(#1738)。
  15. 0.16.23

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.23 让通过一个兼容 OpenAI 端点调度免费模型的人,用 Groq Orpheus 生成英语和沙特阿拉伯语语音。

    FreeLLMAPI 0.16.23 于 2026 年 11 月 9 日发布,为把聊天、图像和音频调用汇到一个兼容 OpenAI 端点的人加入 Groq Orpheus 语音。来源:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.23

    • 用已有的 Groq 密钥,经普通音频路由生成英语和沙特阿拉伯语语音。
    • 返回经过校验的 PCM16 WAV,整理 Groq 流式 WAV 头,并保留故障转移和限速退避。
    • 标准 OpenAI 语音名称映射到已测试的原生默认语音,也接受明确指定的原生语音。
    • Premium 收到实时签名目录;Free 在不变的 30 天等待后收到新条目。
    • Groq 组织管理员必须接受每个 Orpheus 模型的条款。TwelveLabs 和 Mistral 本版不变。
  16. 0.16.22

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.22 于 2026-11-08 发布。本次更新加入 Rime 语音模型和转写,并带上 Varosity 的媒体适配器。

    FreeLLMAPI 0.16.22 于 2026-11-08 发布。本次更新加入 Rime 语音模型和转写,并带上 Varosity 的媒体适配器。

    • 新增 Rime Coda v2 和 Mist v3 的文本转语音,以及 Rime 转写。
    • Rime 进入加密密钥管理和导入。不需要项目 ID。模型通过托管目录提供。
    • 包含 Varosity 已经合并的图像、语音和音乐、以及视频适配器。
    • 包含近期的 APINex 签到恢复、网页搜索就绪和缓存诊断,以及 Waterfall 入门流程的修复。
    • Rime 文本转语音目前返回 WAV。
    • 转写接受 PCM16 WAV,单声道或立体声,8–48 kHz,英语或西班牙语,不需要外部 ffmpeg。
    • 不包含 Prism 实时功能。
    • 目录更新增加 18 个已测试条目,并停用两条已退役的 BeatAPI JEV 路由,净增 16 个启用项。
    • 提供商额度是共用的,不是按模型另给一份。
    • 现有目录权限不变:Premium 得到当前目录;Free 在 30 天后得到每一项新增。
    • Mistral 的目录条目不变。
  17. 0.16.21

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.21 于 2026-11-07 发布。本次更新新增了经过测试的 FreeInference DeepSeek V4 Flash 聊天与流式支持。

    • 新增经过测试的 FreeInference DeepSeek V4 Flash 聊天与流式支持,并保留精确的响应模型校验。
    • 升级以使用新的 DeepSeek catalog 路由。
    • 签名 catalog 交付、本地模型偏好以及 30 天 Free 和 Premium 逐步推送保持不变。
    • Catalog 发布在发布验证后单独处理。
    • 包含 PR #1834。
  18. 0.16.20

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeModel 个人 BYOK 对话,签名目录新增 24 条已启用路由。

    FreeLLMAPI 0.16.20 于 2026-11-06 发布。本次更新加入 FreeModel 个人 BYOK 对话,以及 24 条已启用的目录路由。

    • 在提供商选择、环境变量和密钥导入中加入 FreeModel。
    • 支持文字对话和流式输出,带严格的模型身份检查、空响应保护和常规 HTTP 与退避故障转移。
    • 跟踪该提供商的共享账户额度,不从速率限制响应头推算月度余额。
    • 保持签名目录的投递规则:Premium 立即收到新条目;Free 在既有 30 天窗口之后收到。新的模型种子不能绕过这道门。
    • FreeModel 需要你自己的 API 密钥和 Telegram 验证。其宣传的每月 100 万 token 额度在各模型之间共享。只想免费使用时不要给钱包充值;充值后的钱包在额度用完后可能被扣费。在共享或商业网关使用前请阅读提供商条款;这次集成并不构成转售或竞争服务的许可。
    • 另一次目录更新加入 29 条已测试的 FreeModel 路由和 2 条 Kenari 路由,停用 7 条已退役路由但不删除,已启用目录净增 24。随附适配器上,29 条 FreeModel 路由的文字和流式输出均通过;未经测试的工具和视觉能力不作宣传。
    • 验证:22 项针对 FreeModel 适配器的测试,以及提供商密钥格式覆盖;完整的服务端、CLI 和客户端回归,以及构建、lint 和迁移往返;签名目录的结构、导入和幂等检查;第 0 天、第 29 天和第 30 天的门禁未改。

    来源:https://github.com/tashfeenahmed/freellmapi/releases/tag/v0.16.20(提供商集成 pull request 1830)。

  19. 0.16.19

    展开

    OpenNavo 编辑 Homebrew 收录

    兼容性macOS 12 起

    FreeLLMAPI 0.16.19 于 2026-11-04 发布,新增对 FreeInference 聊天和嵌入的支持。

    • FreeInference 现已加入提供方选择器和密钥导入流程。
    • 支持经过测试的 Qwen3.6 35B 聊天(包括流式输出),以及采用 1,024 维向量的 BGE-M3 嵌入。
    • 会校验上游模型标识和嵌入维度,同时保留正常的路由错误和故障转移。
    • FreeInference 的公开模型列表不再作为 API 密钥校验依据。密钥健康状态由真实请求确定。
    • 签名目录新增两项,且不移除任何现有模型。Premium 会立即收到新目录;Free 仍按未改动的 30 天发布门槛获得这些条目。
    • FreeInference 提供共享的每日研究额度。它会记录提示词和响应用于研究,因此不要发送敏感数据。
  20. 0.16.18

    展开

    GitHub Release

    NovAI image generation

    • Add the tested free cogview-3-flash image route to the existing NovAI provider. Use your NovAI API key with /v1/images/generations for one 1024×1024 image.
    • Validate downloaded image data and return PNG base64, including when the upstream labels JPEG bytes as PNG. Output downloads never receive the API key and are restricted to approved HTTPS hosts.
    • Package the native image decoder in desktop builds; unsupported installations fail only this image route and retain existing media failover.

    Catalog update

    The separately published catalog adds six tested routes and disables five retired routes: net +1 enabled entry. It also corrects Router9 allowance descriptions and Cohere embedding input limits.

    Premium receives the live catalog immediately. Free retains the existing 30-day release delay. No release-gating code or existing model ages were changed. Catalog data is published to Oracle, not embedded in this release.

    Validation

    • 79 targeted NovAI tests and a successful live image-generation test.
    • Full workspace tests: 5,470 passed; seven existing skips.
    • Desktop tests: 79 passed; one existing skip.
    • Workspace builds, lint, migrations and desktop bundles passed locally.
    • Signed catalog dry-runs verify Premium/Free routing and the day-0/day-29/day-30 boundary.

    PR: https://github.com/tashfeenahmed/freellmapi/pull/1814

  21. 0.16.17

    展开

    GitHub Release

    NovAI daily-free chat support

    • Add chat and streaming support for GLM 4.6V Flash and GLM 4.1V Thinking Flash through the existing NovAI provider.
    • Normalize leading reasoning and nested answer envelopes while preserving streaming, tool deltas and retry information.
    • Reject mismatched model identities so the router can fail over safely.
    • Keep existing NovAI video support and the Premium-immediate / Free-after-30-days catalog policy unchanged.

    The separately published catalog enables the two NovAI chat entries and disables three retired onomeo free routes: codestral-latest, ministral-14b-latest and qwen-flash-character. Net enabled change: -1. Existing rows are retained.

    NovAI free models share a daily allowance. GLM 4.6V Flash passed live adapter checks in both modes. The thinking model returned the correct answers and identity in both modes; final formatting normalization is regression-tested, but a post-fix live repeat was blocked by the account's daily quota.

    Upgrade desktop or pull the stable Docker image to use the new adapter.

  22. 0.16.16

    展开

    GitHub Release

    v0.16.16 — Audio and video catalog expansion

    Audio and video catalog support

    This release includes the adapters for the November 1 catalog update: 13 tested media additions, with three duplicate listings disabled (net +10 enabled entries).

    • NovAI: daily-free CogVideoX Flash video generation, with asynchronous polling, cancellation, safe media downloads and normal routing/failover.
    • Free.ai: MeloTTS, Chatterbox, Qwen3 TTS and VibeVoice speech; Whisper uploaded-file transcription.
    • OVH: keyless English, German, Spanish and Italian speech, plus Whisper transcription support.
    • Mistral: Voxtral Mini TTS 3.0 speech.
    • Google: Gemini 3.5 Transcribe through the Interactions API.
    • ElectronHub: support for the catalog's GPT-4o Mini TTS route through the existing speech adapter.
    • Cohere: transcription routing and correct classification of its existing speech-recognition entry.

    Fixes

    • Correct Cloudflare Aura speech payloads and binary audio handling, MeloTTS WAV labeling, and Nova 3 transcription parsing.
    • Keep dedicated transcription routes enabled when their legacy duplicate audio listings are disabled. Local user disables remain respected.

    Rollout

    The signed Oracle catalog is already published as 2026.11.01.fbb00f16: 1,221 total entries, 737 enabled, 57 active providers. Premium receives new catalog entries immediately; Free receives them through the existing 30-day promotion. This release does not bypass that gate or seed new models locally.

    Update the app, configure the relevant provider keys, and sync the catalog. Provider allowances are shared and upstream availability/rate limits still apply. NovAI's free output is watermarked.

    Verification

    All 13 additions passed live application-adapter tests. The adapter PR passed Node 20/22/24 CI and Linux amd64/arm64 Docker builds. Local regression checks passed 4,811 server tests, 178 CLI tests and 436 client tests, plus build, lint, migration and localization checks. Oracle signatures and the unchanged day-0/day-29/day-30 promotion behavior were verified.

    Adapter changes: https://github.com/tashfeenahmed/freellmapi/pull/1797

  23. 0.16.15

    展开

    GitHub Release

    v0.16.15 — nabh.cloud support

    New provider: nabh.cloud

    • Adds built-in nabh.cloud support for normal and streaming chat completions, with API-key setup in the dashboard and environment-file imports.
    • Uses a read-only authenticated model-list request for key health checks, without consuming inference credits.
    • Shares account-level quota across routes, preserves upstream errors for normal fallback, and rejects mismatched response model identities.

    GPT OSS 20B is the tested Free-plan route. Model availability is delivered through the signed live catalog, not bundled model seeds. Premium users receive catalog additions immediately; Free users retain the existing 30-day delay.

    Upgrade

    Install the desktop package for your platform below, or pull ghcr.io/tashfeenahmed/freellmapi:v0.16.15 (also published as latest) for stable Docker. Bring your own nabh.cloud API key.

    Validation

    The adapter passed live normal and streaming tests. Node 20/22/24 CI, both Docker architecture checks, all four desktop packaging checks and macOS updater-metadata validation passed before the release tag. Tagged macOS builds additionally run signing, notarization and packaged-backend smoke checks before publication.


    ⭐ Like the free router? Go Premium — the live signed catalog, $19/yr, cancel anytime.

  24. 0.16.14

    展开

    GitHub Release

    v0.16.14 — JamAI and macOS release fix

    Includes JamAI Base chat and embeddings with separate PAT and Project ID setup fields, plus the Intel macOS release fix.

    Fixes the Apple CloudKit connection failure documented in actions/runner-images#14924 by using HTTPS over TCP on signed Intel GitHub runners. Notarization, ticket validation, Gatekeeper checks, native-backend smoke tests, and update-metadata verification remain required.

    Validation before merge: all Node 20/22/24 CI jobs, both Docker architectures, all four desktop builds and combined Mac metadata checks passed. Full local server, CLI, dashboard and desktop regression suites, lint, builds and 60-locale checks passed.

    The catalog is already published with 15 additions and 2 disabled routes (net +13). This release does not change catalog data or the existing Premium-immediate / Free-after-30-days policy. Update the app to use the new JamAI adapter.

    This release supersedes the stalled v0.16.13 desktop draft; existing tags are unchanged.

    PRs: https://github.com/tashfeenahmed/freellmapi/pull/1764 and https://github.com/tashfeenahmed/freellmapi/pull/1767