wallaby-ai-provider
A third-party Vercel AI SDK provider that integrates the Kimi K3 model into the AI SDK ecosystem. Once installed, Kimi K3 can be called in any environment where the AI SDK runs.
Timeline (last 90 days)
wallaby-ai-provider is a third-party npm package that lets users call Kimi K3 in the Vercel AI SDK.
wallaby-ai-provider is developed by Wallaby Token, and its endpoint points to Wallaby Token's own service, so the package is a commercial-purpose integration rather than a community-neutral open-source project.
Using wallaby-ai-provider to call Kimi K3 requires a Wallaby Token account and API Key, as well as Node.js 18 or higher, and installation via npm install wallaby-ai-provider ai.
One real execution of a generateText call using wallaby('kimi-k3') consumed 92 input tokens and 117 output tokens.
One call using streamText consumed 103 input tokens and 1859 output tokens, costing $0.0254, calculated as (103 × $2.70 + 1859 × $13.50) ÷ 1,000,000.
In the Vercel AI SDK, the model ID must be exactly kimi-k3, otherwise a 'model not found' error occurs.
When integrating in Next.js, WALLABY_API_KEY should be stored in the .env.local file, and after modifying environment variables, next dev must be restarted for the change to take effect.
All Facts (7)
In the Vercel AI SDK, the model ID must be exactly kimi-k3, otherwise a 'model not found' error occurs.
50%UnverifiedWhen integrating in Next.js, WALLABY_API_KEY should be stored in the .env.local file, and after modifying environment variables, next dev must be restarted for the change to take effect.
50%UnverifiedUsing wallaby-ai-provider to call Kimi K3 requires a Wallaby Token account and API Key, as well as Node.js 18 or higher, and installation via npm install wallaby-ai-provider ai.
50%Unverifiedwallaby-ai-provider is a third-party npm package that lets users call Kimi K3 in the Vercel AI SDK.
50%Unverifiedwallaby-ai-provider is developed by Wallaby Token, and its endpoint points to Wallaby Token's own service, so the package is a commercial-purpose integration rather than a community-neutral open-source project.
50%UnverifiedOne real execution of a generateText call using wallaby('kimi-k3') consumed 92 input tokens and 117 output tokens.
50%UnverifiedOne call using streamText consumed 103 input tokens and 1859 output tokens, costing $0.0254, calculated as (103 × $2.70 + 1859 × $13.50) ÷ 1,000,000.
50%