# Purveyors.io

> Coffee intelligence platform for public catalog discovery, market analytics, developer integrations, and account-linked roasting workflows.

## Public Pages

- [Market Analytics](https://www.purveyors.io/analytics): Public market-intelligence surface for origin price trends, processing mix, origin price ranges, and gated Parchment Intelligence modules.
- [Coffee Catalog](https://www.purveyors.io/catalog): Public catalog for normalized green coffee listings with origin, processing, pricing, and availability data.
- [Cherry Evals](https://www.purveyors.io/evals): Domain benchmarks for green coffee, sensory analysis, sourcing, and roasting.
- [PV-Microlot: Agentic Coffee Specialist Benchmark V1 findings](https://www.purveyors.io/evals/coffeebench-v1): Published system benchmark with 1,200 absolute evaluations and all 1,800 pairwise ballots. Harnessed systems beat Raw, but Purveyors did not clearly beat Pi and Parchment added no measured pairwise lift. The report also explains why V1 cannot attribute the full Raw gap to retrieval alone.
- [Parchment API](https://www.purveyors.io/api): Product overview for the API, access tiers, and Console entry points.
- [Parchment Console](https://www.purveyors.io/api-dashboard): Authenticated Console for API keys, usage, and billing.
- [Developer Docs](https://www.purveyors.io/docs): Product and CLI documentation. Generated API reference lives at https://api.purveyors.io/docs.
- [Blog](https://www.purveyors.io/blog): Coffee intelligence, product direction, and platform updates.
- [Purveyors Fieldnotes](https://www.purveyors.io/fieldnotes): Short takes on coffee, AI, and emerging technology. Interesting connections, coffees to explore, and starting points for your next experiment. Account-backed waitlist; weekly delivery is not live yet.
- [PurveyorsBot](https://www.purveyors.io/bot): Public crawler identity, request policy, data use, and opt-out instructions for website operators.

## API and Platform

- [Generated Parchment API reference](https://api.purveyors.io/docs): Canonical OpenAPI/Scalar documentation for HTTP integrations.
- [Parchment API product page](https://www.purveyors.io/api): Access tiers and Console entry points.
- /api/* routes on this host are private BFF routes for the first-party app and Console. They are not the public API contract.

## Blog Posts

- [Coffee growth is pulling closer to origin](https://www.purveyors.io/blog/market-brief-002): This week’s market stayed quiet while origin access, recovery, and support for producing communities moved to the foreground.
- [What Should an Organization Refuse to Build?](https://www.purveyors.io/blog/what-should-an-organization-refuse-to-build): Strategy becomes useful when it gives different specialists a shared direction, earns conviction, and makes its tradeoffs consequential.
- [When Does More Context Make an AI Worse?](https://www.purveyors.io/blog/when-more-context-makes-ai-worse): Agents should support fluid conversation and structured workspaces without making users understand the context machinery that keeps either one reliable.
- [When Your Moat Becomes a Prison](https://www.purveyors.io/blog/which-moats-survive-ai-economy): Kodak, Nokia, and BlackBerry show how durable advantages can constrain adaptation when markets change.
- [What a New Hire Learns That Enterprise AI Still Misses](https://www.purveyors.io/blog/enterprise-second-brains-are-not-knowledge-bases): The right model for enterprise AI is not 'better search.' It is compressed onboarding. A real enterprise second brain teaches an agent what a strong new hire learns in the first six months, then turns that knowledge into reliable action.
- [Is Co-Fermentation Cheating? Wrong Question.](https://www.purveyors.io/blog/co-fermentation-wrong-question): The specialty coffee world is split on whether co-fermented coffee belongs in competition. The more useful question is what processing transparency should require from everyone else.
- [The Codebase Is the Onboarding Manual](https://www.purveyors.io/blog/building-product-philosophy-into-codebase): If agents are blank slate coworkers, product strategy can no longer live in Slack, Notion, or someone else's head. It has to become part of the execution environment.
- [Sycophancy Is the Last Hard Problem in AI-Assisted Software Development](https://www.purveyors.io/blog/sycophancy-is-the-last-hard-problem): RLHF-trained coding agents do not just make mistakes. They silently implement the wrong thing, accruing alignment debt that passes tests and leaves the codebase worse off.
- [Why Does Enterprise AI Cost More and Deliver Less?](https://www.purveyors.io/blog/why-does-enterprise-ai-cost-more): Microsoft shipped an enterprise Anthropic integration in weeks. They clearly can deliver frontier AI. So why does Copilot for 365 feel like a downgrade from a $20 subscription?
- [Who Profits When Coffee Data Stays Scarce?](https://www.purveyors.io/blog/who-profits-when-coffee-data-stays-scarce): The specialty coffee industry solved price transparency. The harder problem, product metadata, remains relationship-gated. Across 35 suppliers, disclosure ranges from 1 to 13 attributes per listing. That gap is not an accident; it is an economic structure with beneficiaries.
- [Inference Is in the Name](https://www.purveyors.io/blog/inference-is-in-the-name): If your workflow has no unknowns, runtime LLM inference is often the wrong architecture. Build deterministic cores, then add AI where adaptation actually matters.
- [Beyond the Coffee Belt](https://www.purveyors.io/blog/beyond-the-coffee-belt): The coffee belt is a useful heuristic, but the tails of the distribution reveal where quality, risk, and long-term value are really heading.
- [Benchmark Leaders, Agentic Laggards](https://www.purveyors.io/blog/benchmark-leaders-agentic-laggards): The AI leaderboard tells you which model reasons best in isolation. It tells you almost nothing about which model completes real work.
- [What the Fair Use Conversation Is Missing About LLM Data Extraction](https://www.purveyors.io/blog/llm-fair-use-data-extraction): The AI copyright debate focuses on training data. But the more commercially relevant question might be extraction, and the legal framework for it may already exist.
- [The Real AI Moats Aren't Software](https://www.purveyors.io/blog/ai-moats-arent-software): Citrini's viral doomsday report and Citadel's rebuttal are both missing the interesting question. If software isn't the moat anymore, what is?
- [What I Learned Running OpenClaw as a Solo Dev for Two Weeks](https://www.purveyors.io/blog/two-weeks-with-ai-co-developer): Agentic harnesses solve the orchestration problem. The models are the bottleneck. Here's what actually works after 43 PRs and a zettelkasten full of operational data.
- [Building a Coffee Data Pipeline from Scratch](https://www.purveyors.io/blog/building-a-coffee-data-pipeline): The most interesting thing about the purveyors data pipeline isn't the scraping. It's the recursive feedback loop, and what it reveals about directing AI agents.
- [What is Purveyors?](https://www.purveyors.io/blog/what-is-purveyors): The green coffee supply chain is built for commercial buyers. If you're a hobbyist, you're an afterthought. That's the problem purveyors.io exists to solve.

## Data and Workflows

- Normalized supplier listings with origin, legacy processing labels, structured process transparency, grade, pricing, availability metadata, and opt-in proof summaries
- Beta catalog similarity matching for likely-same-bean and similar-profile research with cautious confidence labels
- Daily pricing and availability snapshots for catalog and analytics surfaces
- Inventory, roast, sales, tasting, chat, and workspace workflows in the web app
- Shared CLI-backed tooling for terminal and agent workflows