{"slug": "i-run-a-radar-that-finds-free-llm-endpoints-and-auto-adopts-the-good-ones-behind", "title": "I run a 'radar' that finds free LLM endpoints and auto-adopts the good ones — behind a five-part gate so it can't adopt junk", "summary": "A developer built model-radar, an open-source tool that automatically discovers free LLM endpoints and adopts qualifying models into an agent stack, but only after they pass a fixed five-stage deterministic gate. The gate tests agent-critical behaviors such as tool-calling, full-prompt retention, and stability on real tasks, rejecting any candidate that fails a single stage to prevent silent failures from cheaper models.", "body_md": "Free LLM endpoints appear and vanish constantly — a provider opens a generous\n\ntier, a new open model shows up on someone's gateway, a promo runs for a month.\n\nChasing them by hand is a losing game. So I built a small radar\n\n([model-radar](https://github.com/casareanderson/model-radar)) that watches for\n\nfree models and, when one is genuinely good enough, **adopts it automatically** into\n\nmy agent stack — but only after it clears a gate strict enough that it has never\n\nonce adopted something that embarrassed me.\n\nThe auto-adoption is the flashy bit. The *gate* is the bit worth your time, because\n\n\"just use the free model\" is advice that will quietly wreck an agent pipeline.\n\nThe appeal is obvious: a lot of agent work — summarise, triage, draft, classify —\n\ndoesn't need a frontier model, and paying frontier prices for it is silly. A free\n\ntier that's \"good enough\" for the boring 90% is real money saved.\n\nThe catch is that \"free model\" and \"works in an agent\" are almost unrelated\n\nproperties. A model can score beautifully on a leaderboard and be **useless for agentic work**, because agent work needs things benchmarks don't measure:\n\nSo a radar that adopted models on *quality scores* would be a machine for\n\nintroducing silent failures. The whole design problem is: **how do you let a process adopt something automatically without letting it adopt something bad?**\n\nThe radar can *find* candidates freely. It can *promote* one only by passing every\n\nstage of a fixed, deterministic gate — and if any stage fails, it's rejected, full\n\nstop. No score, no averaging, no \"it's mostly fine.\" The gate checks, in order,\n\nthings a leaderboard never will:\n\nOnly a model that clears all five gets adopted. Everything is **measured on real tasks and written down**, so adoption is a reproducible decision I can audit later,\n\nThe principle underneath is the same one my whole estate runs on: **let the\nautomated part gather and propose freely; let it *act* only through a gate that a\nbad candidate cannot argue its way past.** A guardrail written as a suggestion in a\n\nprompt is decoration. A guardrail written as a pass/fail function is a wall.\n\nThe failure I designed the whole thing to avoid is the one that gets everyone else:\n\na cheaper model quietly swapped in that *looks* like it's working — emits calls,\n\nreturns fluent text — while ignoring tool results or truncating context, producing\n\nconfident garbage that no error ever flags. The gate exists specifically to catch\n\nthe model that fails *silently*, because that's the only kind of failure that\n\nsurvives long enough to hurt you.\n\n[model-radar is on GitHub](https://github.com/casareanderson/model-radar). Even if\n\nyou never auto-adopt anything, the transferable idea is the gate: **before you route real work to a free model, test it on the thing agents actually need — can it\ntool-call and use the result, does it keep your whole prompt, is it stable — and\nmake that test a hard pass/fail, not a score.** Cheap intelligence is worth\n\n*Notes from someone who wanted a coffee-a-month agent bill and refused to pay for it in silent failures, so built a radar that adopts free models automatically and a\ngate strict enough to trust it.*\n\n🤖 *Drafted with AI assistance from my own homelab notes, logs and repos, then reviewed and edited before publishing.*", "url": "https://wpnews.pro/news/i-run-a-radar-that-finds-free-llm-endpoints-and-auto-adopts-the-good-ones-behind", "canonical_source": "https://dev.to/c1-anderson/i-run-a-radar-that-finds-free-llm-endpoints-and-auto-adopts-the-good-ones-behind-a-five-part-3bja", "published_at": "2026-09-16 08:00:03+00:00", "updated_at": "2026-09-16 08:13:31.209987+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "large-language-models", "ai-infrastructure", "developer-tools"], "entities": ["model-radar", "GitHub"], "alternates": {"html": "https://wpnews.pro/news/i-run-a-radar-that-finds-free-llm-endpoints-and-auto-adopts-the-good-ones-behind", "markdown": "https://wpnews.pro/news/i-run-a-radar-that-finds-free-llm-endpoints-and-auto-adopts-the-good-ones-behind.md", "text": "https://wpnews.pro/news/i-run-a-radar-that-finds-free-llm-endpoints-and-auto-adopts-the-good-ones-behind.txt", "jsonld": "https://wpnews.pro/news/i-run-a-radar-that-finds-free-llm-endpoints-and-auto-adopts-the-good-ones-behind.jsonld"}}