{"slug": "seeking-advice-feedback-building-a-local-ai-to-setup-multiple-trading-agents", "title": "Seeking advice/feedback: Building a local AI to setup multiple trading agents", "summary": "A community forum reply recommends running multiple trading agents against a single vLLM server, which provides continuous batching, chunked prefill, prefix caching and an OpenAI-compatible API so each agent acts as a client. The reply also advises keeping the order-execution tool out of the LLM's tool set until paper trading is complete, then placing a deterministic code-level gate in front of it with an instrument allowlist, size limits and a human approval step. The author discloses maintaining agent-airlock, an in-process Python decorator that validates tool-call arguments and enforces an allow/deny tool policy with an approver hook that refuses calls when no approver responds or the request times out, noting that an empty allowlist permits every tool and that no Slack transport is built in.", "body_md": "I’ll leave the hardware and model picks to others. Two thoughts on the stack.\n\nFor many agents sharing one model, vLLM’s server fits: it does “Continuous batching of incoming requests, chunked prefill, prefix caching” and runs an “OpenAI-compatible API server”, so each agent is just a client.\n\nOn trade execution, I feel the order tool should stay out of the LLM’s tool set until paper trading is done. When it goes in, put a deterministic gate in front of it in code, not in the prompt: an instrument allowlist, size limits, and a human approval step.\n\nDisclosure: I maintain agent-airlock, an in-process Python decorator that validates tool-call arguments and applies an allow/deny tool policy, with an approver hook that refuses the call when no approver answers or it times out. Limits: an empty allowlist allows every tool, the tool call waits for approval, and there is no Slack transport built in, so you would wire that yourself.\n\n[https://github.com/sattyamjjain/agent-airlock](https://github.com/sattyamjjain/agent-airlock)", "url": "https://wpnews.pro/news/seeking-advice-feedback-building-a-local-ai-to-setup-multiple-trading-agents", "canonical_source": "https://discuss.huggingface.co/t/seeking-advice-feedback-building-a-local-ai-to-setup-multiple-trading-agents/191580#post_2", "published_at": "2026-11-09 04:19:11+00:00", "updated_at": "2026-11-09 04:22:04.961927+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "large-language-models"], "entities": ["vLLM", "agent-airlock", "OpenAI", "GitHub"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/seeking-advice-feedback-building-a-local-ai-to-setup-multiple-trading-agents", "markdown": "https://wpnews.pro/news/seeking-advice-feedback-building-a-local-ai-to-setup-multiple-trading-agents.md", "text": "https://wpnews.pro/news/seeking-advice-feedback-building-a-local-ai-to-setup-multiple-trading-agents.txt", "jsonld": "https://wpnews.pro/news/seeking-advice-feedback-building-a-local-ai-to-setup-multiple-trading-agents.jsonld"}}