<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>FlintAPI Blog</title>
  <subtitle>Verified trade data and AI infrastructure, written down.</subtitle>
  <link href="https://flintapi.ai/feed.xml" rel="self"/>
  <link href="https://flintapi.ai/blog"/>
  <id>https://flintapi.ai/</id>
  <updated>2026-11-09T00:00:00Z</updated>
  <entry>
    <title>Import concentration, four cases: Vietnam phones, Indonesia fuel, Thailand chips, Germany cars</title>
    <link href="https://flintapi.ai/blog/2026-11-09-import-concentration-four-cases"/>
    <id>https://flintapi.ai/blog/2026-11-09-import-concentration-four-cases</id>
    <updated>2026-11-09T00:00:00Z</updated>
    <published>2026-11-09T00:00:00Z</published>
    <summary>Four 2023 import markets from UN Comtrade. Named-supplier concentration runs from 85% down to 10% - and in two of the four markets the figures cannot even name the largest supplier, which is itself worth knowing before you write an email.</summary>
    <category term="tradedata"/>
    <category term="importstatistics"/>
    <category term="uncomtrade"/>
    <category term="marketentry"/>
  </entry>
  <entry>
    <title>How to size an import market with UN Comtrade (worked example: Indonesia, HS 2106)</title>
    <link href="https://flintapi.ai/blog/2026-10-30-how-to-size-an-import-market-with-comtrade"/>
    <id>https://flintapi.ai/blog/2026-10-30-how-to-size-an-import-market-with-comtrade</id>
    <updated>2026-10-30T00:00:00Z</updated>
    <published>2026-10-30T00:00:00Z</published>
    <summary>A step-by-step method to size an import market from the free UN Comtrade API, with a full worked example (Indonesia, HS 2106, 2023) whose every figure you can re-check in one call.</summary>
    <category term="marketresearch"/>
    <category term="tradedata"/>
    <category term="comtrade"/>
    <category term="marketentry"/>
  </entry>
  <entry>
    <title>Reading import partner concentration before you pick a market</title>
    <link href="https://flintapi.ai/blog/2026-10-30-reading-import-partner-concentration"/>
    <id>https://flintapi.ai/blog/2026-10-30-reading-import-partner-concentration</id>
    <updated>2026-10-30T00:00:00Z</updated>
    <published>2026-10-30T00:00:00Z</published>
    <summary>Market size is the easy number. Partner concentration tells you whether a new entrant can actually get in — worked example with Indonesia versus Vietnam on HS 2106 (2023).</summary>
    <category term="marketresearch"/>
    <category term="tradedata"/>
    <category term="marketentry"/>
    <category term="comtrade"/>
  </entry>
  <entry>
    <title>Sample deliverable: deciding between a price increase and a cheaper tier</title>
    <link href="https://flintapi.ai/blog/2026-10-28-sample-consulting-report"/>
    <id>https://flintapi.ai/blog/2026-10-28-sample-consulting-report</id>
    <updated>2026-10-28T00:00:00Z</updated>
    <published>2026-10-28T00:00:00Z</published>
    <summary>A real, unedited consulting output — the full report our pipeline produced for a pricing decision question on a 30k MRR B2B SaaS, with honest timing and cost notes.</summary>
    <category term="consulting"/>
    <category term="sample"/>
    <category term="deliverables"/>
  </entry>
  <entry>
    <title>The character limit that silently killed our AI replies</title>
    <link href="https://flintapi.ai/blog/character-limit-silently-killed-replies"/>
    <id>https://flintapi.ai/blog/character-limit-silently-killed-replies</id>
    <updated>2026-10-28T00:00:00Z</updated>
    <published>2026-10-28T00:00:00Z</published>
    <summary>We pasted our whole knowledge base into the prompt. The endpoint refused it. Our code answered with a template instead.</summary>
    <category term="ai"/>
    <category term="debugging"/>
    <category term="llm"/>
    <category term="api"/>
  </entry>
  <entry>
    <title>What Chinese LLM APIs actually cost in 2026 (measured per 1M tokens)</title>
    <link href="https://flintapi.ai/blog/chinese-llm-api-pricing-2026"/>
    <id>https://flintapi.ai/blog/chinese-llm-api-pricing-2026</id>
    <updated>2026-10-28T00:00:00Z</updated>
    <published>2026-10-28T00:00:00Z</published>
    <summary>A dated price table for twelve Chinese models, why list price is not your bill, and the one line of code that shows your real number.</summary>
    <category term="ai"/>
    <category term="api"/>
    <category term="llm"/>
    <category term="pricing"/>
  </entry>
  <entry>
    <title>Why our health check said &quot;up&quot; while the model was gone</title>
    <link href="https://flintapi.ai/blog/health-check-said-up-model-was-gone"/>
    <id>https://flintapi.ai/blog/health-check-said-up-model-was-gone</id>
    <updated>2026-10-28T00:00:00Z</updated>
    <published>2026-10-28T00:00:00Z</published>
    <summary>Our status page was green for weeks while generations were failing. The check was asking the wrong question.</summary>
    <category term="devops"/>
    <category term="api"/>
    <category term="monitoring"/>
    <category term="llm"/>
  </entry>
  <entry>
    <title>Selling tokens below cost — what a margin audit found in our own gateway</title>
    <link href="https://flintapi.ai/blog/selling-tokens-below-cost"/>
    <id>https://flintapi.ai/blog/selling-tokens-below-cost</id>
    <updated>2026-10-28T00:00:00Z</updated>
    <published>2026-10-28T00:00:00Z</published>
    <summary>We audited our own rate card against upstream list prices and found five models priced below cost, one of them by about two to one. Here is the arithmetic, the mistake that caused it, and the checks we now run.</summary>
    <category term="llm"/>
    <category term="pricing"/>
    <category term="engineering"/>
    <category term="operations"/>
  </entry>
  <entry>
    <title>What actually broke when I put four Chinese LLM providers behind one API</title>
    <link href="https://flintapi.ai/blog/what-breaks-chinese-llm-gateway"/>
    <id>https://flintapi.ai/blog/what-breaks-chinese-llm-gateway</id>
    <updated>2026-10-28T00:00:00Z</updated>
    <published>2026-10-28T00:00:00Z</published>
    <summary>Timeouts, empty 200s, and why a per-attempt deadline beats a bigger one.</summary>
    <category term="ai"/>
    <category term="api"/>
    <category term="llm"/>
    <category term="webdev"/>
  </entry>
  <entry>
    <title>What we tell users when 9 of our 13 models are down</title>
    <link href="https://flintapi.ai/blog/what-we-tell-users-when-models-are-down"/>
    <id>https://flintapi.ai/blog/what-we-tell-users-when-models-are-down</id>
    <updated>2026-10-28T00:00:00Z</updated>
    <published>2026-10-28T00:00:00Z</published>
    <summary>A gateway&apos;s worst failure is not a crash, it is a model that answers 503 for a week while the dashboard says &quot;up&quot;. Here is the contract we now publish, the mistake that caused a silent outage, and the exact 503 body we return when a model is unavailable.</summary>
    <category term="llm"/>
    <category term="operations"/>
    <category term="reliability"/>
    <category term="api"/>
  </entry>
  <entry>
    <title>The Chinese LLM Landscape in 2026: A Practical Guide</title>
    <link href="https://flintapi.ai/blog/2026-08-20-chinese-llm-landscape-2026"/>
    <id>https://flintapi.ai/blog/2026-08-20-chinese-llm-landscape-2026</id>
    <updated>2026-08-20T00:00:00Z</updated>
    <published>2026-08-20T00:00:00Z</published>
    <summary>DeepSeek, Qwen, GLM, Kimi — what each is best at, and how to access them all from one API.</summary>
    <category term="models"/>
    <category term="overview"/>
    <category term="china"/>
  </entry>
  <entry>
    <title>Build a Multimodal App in an Afternoon with FlintAPI</title>
    <link href="https://flintapi.ai/blog/2026-08-17-build-multimodal-app-flintapi"/>
    <id>https://flintapi.ai/blog/2026-08-17-build-multimodal-app-flintapi</id>
    <updated>2026-08-17T00:00:00Z</updated>
    <published>2026-08-17T00:00:00Z</published>
    <summary>Image, video, audio, and text — one API key, one endpoint style, four modalities.</summary>
    <category term="multimodal"/>
    <category term="tutorial"/>
    <category term="vision"/>
  </entry>
  <entry>
    <title>How Flint-Smart Routing Can Cut Your AI Bill</title>
    <link href="https://flintapi.ai/blog/2026-08-14-flint-smart-routing-cost-savings"/>
    <id>https://flintapi.ai/blog/2026-08-14-flint-smart-routing-cost-savings</id>
    <updated>2026-08-14T00:00:00Z</updated>
    <published>2026-08-14T00:00:00Z</published>
    <summary>Automatic model selection routes every prompt to the cheapest model that does the job well — without you thinking about it.</summary>
    <category term="routing"/>
    <category term="cost"/>
    <category term="tutorial"/>
  </entry>
  <entry>
    <title>DeepSeek V4 Pro on FlintAPI: Quick Start</title>
    <link href="https://flintapi.ai/blog/deepseek-v4-api-guide"/>
    <id>https://flintapi.ai/blog/deepseek-v4-api-guide</id>
    <updated>2026-06-29T00:00:00Z</updated>
    <published>2026-06-29T00:00:00Z</published>
    <summary>Get started with DeepSeek V4 Pro on FlintAPI in under 5 minutes, with a model comparison guide.</summary>
    <category term="tutorial"/>
    <category term="deepseek"/>
    <category term="models"/>
  </entry>
  <entry>
    <title>Why Your AI Stack Needs a Unified API in 2026</title>
    <link href="https://flintapi.ai/blog/why-unified-ai-api"/>
    <id>https://flintapi.ai/blog/why-unified-ai-api</id>
    <updated>2026-06-28T00:00:00Z</updated>
    <published>2026-06-28T00:00:00Z</published>
    <summary>Managing 15 AI provider accounts is a mess. FlintAPI unifies a dozen models behind one API key.</summary>
    <category term="overview"/>
    <category term="api"/>
  </entry>
  <entry>
    <title>Flint-Smart: The Model That Picks the Best Model</title>
    <link href="https://flintapi.ai/blog/flint-smart-routing"/>
    <id>https://flintapi.ai/blog/flint-smart-routing</id>
    <updated>2026-06-26T00:00:00Z</updated>
    <published>2026-06-26T00:00:00Z</published>
    <summary>Stop choosing models manually. Flint-Smart analyzes your prompt and routes to the best model automatically.</summary>
    <category term="routing"/>
    <category term="flintsmart"/>
  </entry>
</feed>
