LightAgent's provider list covers the major Chinese models — Zhipu, DeepSeek, Qwen, Kimi, MiniMax, Baichuan — and with READMEs in 11 languages, the project is clearly reaching a global audience. But there's a gap: for users outside China, signing up with each of these providers requires a Chinese phone number and Alipay. International users who want to try LightAgent with the Chinese models it's built around often can't.
I built AnyLLM, an OpenAI-compatible API gateway with 24+ Chinese and international models behind a single key: DeepSeek V4, Qwen3.7-Plus, GLM-5.2, Kimi K2.5, MiniMax M3 — alongside GPT-5 and Gemini. International cards and crypto accepted, pass-through pricing with zero markup, built-in routing and failover.
Why this fits LightAgent
Zero integration. LightAgent takes model + api_key + base_url — AnyLLM works with exactly that, no adapter needed.
Unblocks the global audience. One AnyLLM key gives international users all the Chinese providers in your list, plus GPT-5 and Gemini in the same connection — a single docs example covers every provider you support.
Multi-agent cost control. LightSwarm delegation and ToT reasoning multiply LLM calls per task. Chinese models at pass-through pricing make larger agent swarms practical — assign cost-efficient models to simple agents, frontier models to reasoning-heavy ones.
Failover aligns with production goals. LightAgent targets production deployments; AnyLLM's built-in provider failover adds resilience with no extra code.
Proposal
A short section in docs/model_providers.md — "Using an OpenAI-compatible gateway (AnyLLM)" — showing the three-line config. Happy to open a PR, and glad to provide a sandbox key for testing.
(Context: I'm the developer of AnyLLM.)
Best,
Leo Bennett
[leo.indiedev@gmail.com]
LightAgent's provider list covers the major Chinese models — Zhipu, DeepSeek, Qwen, Kimi, MiniMax, Baichuan — and with READMEs in 11 languages, the project is clearly reaching a global audience. But there's a gap: for users outside China, signing up with each of these providers requires a Chinese phone number and Alipay. International users who want to try LightAgent with the Chinese models it's built around often can't.
I built AnyLLM, an OpenAI-compatible API gateway with 24+ Chinese and international models behind a single key: DeepSeek V4, Qwen3.7-Plus, GLM-5.2, Kimi K2.5, MiniMax M3 — alongside GPT-5 and Gemini. International cards and crypto accepted, pass-through pricing with zero markup, built-in routing and failover.
Why this fits LightAgent
Zero integration. LightAgent takes model + api_key + base_url — AnyLLM works with exactly that, no adapter needed.
Unblocks the global audience. One AnyLLM key gives international users all the Chinese providers in your list, plus GPT-5 and Gemini in the same connection — a single docs example covers every provider you support.
Multi-agent cost control. LightSwarm delegation and ToT reasoning multiply LLM calls per task. Chinese models at pass-through pricing make larger agent swarms practical — assign cost-efficient models to simple agents, frontier models to reasoning-heavy ones.
Failover aligns with production goals. LightAgent targets production deployments; AnyLLM's built-in provider failover adds resilience with no extra code.
Proposal
A short section in docs/model_providers.md — "Using an OpenAI-compatible gateway (AnyLLM)" — showing the three-line config. Happy to open a PR, and glad to provide a sandbox key for testing.
(Context: I'm the developer of AnyLLM.)
Best,
Leo Bennett
[leo.indiedev@gmail.com]