https://armature.tech/library How coding agents choose which software to install, and what a vendor can do about it. en Thu, 03 Sep 2026 09:00:00 GMT https://armature.tech/library/write-content-an-ai-answer-will-quote https://armature.tech/library/write-content-an-ai-answer-will-quote Thu, 03 Sep 2026 09:00:00 GMT Playbooks Extraction rewards one specific shape: a question heading, then a self-contained answer in 40 to 60 words. Most pages bury the answer in paragraph four. https://armature.tech/library/write-a-quickstart-an-agent-can-follow https://armature.tech/library/write-a-quickstart-an-agent-can-follow Thu, 03 Sep 2026 09:00:00 GMT Playbooks The highest return change most developer tool companies can make. What breaks, why it costs installs, and the exact shape that works. https://armature.tech/library/why-did-the-agent-pick-my-competitor https://armature.tech/library/why-did-the-agent-pick-my-competitor Thu, 03 Sep 2026 09:00:00 GMT Answers Five reasons, ranked by how often they occur in judged sessions. Four of the five are defects you can fix in days, not preferences you have to out-argue. https://armature.tech/library/which-coding-agent-should-i-optimise-for https://armature.tech/library/which-coding-agent-should-i-optimise-for Thu, 03 Sep 2026 09:00:00 GMT Answers They need different work. Codex reads the web in half its decisions, Claude Code in 1.6%. The answer depends on one thing you can find out about your own users. https://armature.tech/library/what-decides-the-pick-the-repository https://armature.tech/library/what-decides-the-pick-the-repository Thu, 03 Sep 2026 09:00:00 GMT From the experiment The same request in a JavaScript and a TypeScript codebase produced different winners. An incumbent beat a written instruction. https://armature.tech/library/what-coding-agents-choose-in-typescript https://armature.tech/library/what-coding-agents-choose-in-typescript Thu, 03 Sep 2026 09:00:00 GMT From the experiment 2,084 judged sessions inside TypeScript codebases. The leader differed from the overall leader in 3 of 15 categories. https://armature.tech/library/what-coding-agents-choose-in-python https://armature.tech/library/what-coding-agents-choose-in-python Thu, 03 Sep 2026 09:00:00 GMT From the experiment 983 judged sessions inside Python codebases. The leader differed from the overall leader in 4 of 7 categories. https://armature.tech/library/what-coding-agents-choose-in-javascript https://armature.tech/library/what-coding-agents-choose-in-javascript Thu, 03 Sep 2026 09:00:00 GMT From the experiment 900 judged sessions inside JavaScript codebases. The leader differed from the overall leader in 2 of 7 categories. https://armature.tech/library/what-coding-agents-build-in-house https://armature.tech/library/what-coding-agents-build-in-house Thu, 03 Sep 2026 09:00:00 GMT From the experiment In five of eighteen categories the most common outcome was hand-written code, not a product. Where agents build rather than adopt. https://armature.tech/library/what-about-the-other-coding-agents https://armature.tech/library/what-about-the-other-coding-agents Thu, 03 Sep 2026 09:00:00 GMT Guides We measured three agents and not the rest. Why those three, what we think transfers to the others, and what we are not willing to claim without data. https://armature.tech/library/voice-agents-coding-agents-playbook https://armature.tech/library/voice-agents-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks Vapi took 24% of 308 judged voice agents sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/time-to-first-successful-call https://armature.tech/library/time-to-first-successful-call Thu, 03 Sep 2026 09:00:00 GMT Playbooks The oldest onboarding metric, with a new reader. An agent does not persist through friction, it switches. What that changes about where you measure. https://armature.tech/library/the-dataset https://armature.tech/library/the-dataset Thu, 03 Sep 2026 09:00:00 GMT From the experiment 235 ranking rows across 18 categories from 5,292 judged sessions, as CSV and JSON, under CC BY 4.0. Cite it, load it, feed it to an agent. https://armature.tech/library/technical-content-marketing-for-ai-search https://armature.tech/library/technical-content-marketing-for-ai-search Thu, 03 Sep 2026 09:00:00 GMT Guides Agents search with long, site-scoped queries and want configuration, limits and pricing units. Most content answers something else. https://armature.tech/library/structured-data-for-developer-tools https://armature.tech/library/structured-data-for-developer-tools Thu, 03 Sep 2026 09:00:00 GMT Glossary Which schema types are worth adding, which are a waste, and the one rule that turns markup from an asset into a manual penalty. https://armature.tech/library/storage-coding-agents-playbook https://armature.tech/library/storage-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks Amazon S3 took 46% of 90 judged file storage sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/share-of-voice https://armature.tech/library/share-of-voice Thu, 03 Sep 2026 09:00:00 GMT Glossary Your brand mentions as a proportion of all brand mentions across a tracked prompt set. Useful against your own history, meaningless as an absolute number. https://armature.tech/library/serverless-coding-agents-playbook https://armature.tech/library/serverless-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks AWS Lambda took 24% of 287 judged serverless compute sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/self-preference https://armature.tech/library/self-preference Thu, 03 Sep 2026 09:00:00 GMT Glossary Whether a coding agent favours its own maker's products. One category showed it, seven did not. What we found and how we controlled for it. https://armature.tech/library/search-coding-agents-playbook https://armature.tech/library/search-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks Postgres Full-Text Search took 17% of 459 judged search sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/sandboxes-coding-agents-playbook https://armature.tech/library/sandboxes-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks E2B took 42% of 299 judged code sandboxes sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/robots-txt-for-ai-crawlers https://armature.tech/library/robots-txt-for-ai-crawlers Thu, 03 Sep 2026 09:00:00 GMT Playbooks Which AI user agents exist, what each one is for, and why blocking the wrong one removes you from the answers instead of protecting you. https://armature.tech/library/readme-and-repository-optimization https://armature.tech/library/readme-and-repository-optimization Thu, 03 Sep 2026 09:00:00 GMT Playbooks A public repository is a first-class artefact for an agent. What to put in it, and the parts that decide whether your code is copied. https://armature.tech/library/prompt-monitoring https://armature.tech/library/prompt-monitoring Thu, 03 Sep 2026 09:00:00 GMT Glossary Sending a fixed list of prompts to AI assistants on a schedule and counting brand mentions. How every AI visibility tool works, and what the method cannot see. https://armature.tech/library/programmatic-seo https://armature.tech/library/programmatic-seo Thu, 03 Sep 2026 09:00:00 GMT Glossary Generating many pages from one data set works when each page carries data only you have. It fails when they are one page renamed. https://armature.tech/library/products-coding-agents-consider-and-never-choose https://armature.tech/library/products-coding-agents-consider-and-never-choose Thu, 03 Sep 2026 09:00:00 GMT From the experiment 67 products reached the shortlist at least 25 times and were adopted zero times. Being rejected is cheaper to fix than being unknown. https://armature.tech/library/product-analytics-coding-agents-playbook https://armature.tech/library/product-analytics-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks PostHog took 53% of 359 judged product analytics sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/pricing-pages-agents-can-read https://armature.tech/library/pricing-pages-agents-can-read Thu, 03 Sep 2026 09:00:00 GMT Playbooks Agents search for pricing more than you would expect, and they want units rather than plan names. Most pricing pages miss that. https://armature.tech/library/persona-prompt https://armature.tech/library/persona-prompt Thu, 03 Sep 2026 09:00:00 GMT Glossary The same request written as four kinds of person. It changed the category leader in 14 of 18 categories, which is the cheapest split. https://armature.tech/library/perf-ci-coding-agents-playbook https://armature.tech/library/perf-ci-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks JMH took 10% of 216 judged performance testing sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/payments-coding-agents-playbook https://armature.tech/library/payments-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks Stripe took 88% of 395 judged payments sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/package-naming-for-coding-agents https://armature.tech/library/package-naming-for-coding-agents Thu, 03 Sep 2026 09:00:00 GMT Playbooks When the product name, the package name and the import name differ, an agent has to guess a mapping. Sometimes it installs somebody else's package. https://armature.tech/library/open-source-as-distribution-for-agents https://armature.tech/library/open-source-as-distribution-for-agents Thu, 03 Sep 2026 09:00:00 GMT Playbooks A public repository was always readable, quotable and copyable. What changed is who copies from it, and how little they forgive. https://armature.tech/library/open-seam https://armature.tech/library/open-seam Thu, 03 Sep 2026 09:00:00 GMT Glossary The thing a test repository is missing, which is what makes it a valid test. A repository that already has your category cannot measure choice. https://armature.tech/library/observability-coding-agents-playbook https://armature.tech/library/observability-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks Sentry took 37% of 360 judged observability sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/model-context-protocol https://armature.tech/library/model-context-protocol Thu, 03 Sep 2026 09:00:00 GMT Glossary The open standard for connecting AI clients to tools and data. What it is, why it changed developer tool distribution, and what it means for discoverability. https://armature.tech/library/mcp-server-discoverability https://armature.tech/library/mcp-server-discoverability Thu, 03 Sep 2026 09:00:00 GMT Playbooks How an agent finds, chooses and keeps using an MCP server. What to put in tool descriptions, why registries are not the main route yet, and what to measure. https://armature.tech/library/mcp-analytics https://armature.tech/library/mcp-analytics Thu, 03 Sep 2026 09:00:00 GMT Guides When a user delegates to an agent, the session happens inside Claude or ChatGPT and your interface never sees it. What to capture. https://armature.tech/library/markdown-twin https://armature.tech/library/markdown-twin Thu, 03 Sep 2026 09:00:00 GMT Glossary The same page served as Markdown at the same address plus .md. Agents read it far better than a rendered site, and almost nobody publishes one. https://armature.tech/library/mail-coding-agents-playbook https://armature.tech/library/mail-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks Resend took 36% of 208 judged email sending sessions. What the numbers say a vendor in this category should do. https://armature.tech/library/llms-txt-guide https://armature.tech/library/llms-txt-guide Thu, 03 Sep 2026 09:00:00 GMT Playbooks About one site in ten has one. It takes an hour. Here is what the file is for, an honest account of what it does and does not do, and a template. https://armature.tech/library/llmo https://armature.tech/library/llmo Thu, 03 Sep 2026 09:00:00 GMT Glossary A third name for the same work as GEO and AEO. What the term adds, what it does not, and the one distinction that is actually load bearing. https://armature.tech/library/judge-model https://armature.tech/library/judge-model Thu, 03 Sep 2026 09:00:00 GMT Glossary The model that reads a session and decides what was chosen. Why it must come from a different family, and the rules that matter. https://armature.tech/library/is-geo-worth-it-for-a-developer-tool https://armature.tech/library/is-geo-worth-it-for-a-developer-tool Thu, 03 Sep 2026 09:00:00 GMT Answers Partly, and the honest answer depends on one number you can measure: how many of your users run an agent that searches the web at all. https://armature.tech/library/install-share https://armature.tech/library/install-share Thu, 03 Sep 2026 09:00:00 GMT Glossary The percentage of sessions in a category where a coding agent picked your product and wrote working integration code. The agent-era equivalent of market share. https://armature.tech/library/in-house-rate https://armature.tech/library/in-house-rate Thu, 03 Sep 2026 09:00:00 GMT Glossary The share of sessions where a coding agent wrote the implementation itself instead of adopting any product. It ranged from 0% to 52% by category. https://armature.tech/library/how-we-measured-this https://armature.tech/library/how-we-measured-this Thu, 03 Sep 2026 09:00:00 GMT Guides The method behind every number here: 51 realistic codebases, 481 frozen prompts, real agents at pinned versions, and a blind judge. https://armature.tech/library/how-to-win-the-enterprise-persona https://armature.tech/library/how-to-win-the-enterprise-persona Thu, 03 Sep 2026 09:00:00 GMT Playbooks Adding procurement language moved the winner to a Microsoft product in four unrelated categories. The mechanism is documentation. https://armature.tech/library/how-to-show-up-in-chatgpt-claude-and-perplexity https://armature.tech/library/how-to-show-up-in-chatgpt-claude-and-perplexity Thu, 03 Sep 2026 09:00:00 GMT Playbooks The three assistants retrieve differently, so the same page performs differently in each. What each one favours, and the work that pays across all three. https://armature.tech/library/how-to-measure-install-share https://armature.tech/library/how-to-measure-install-share Thu, 03 Sep 2026 09:00:00 GMT Playbooks The number that tells you whether coding agents install your product. How to measure it yourself in a weekend, and the traps that make a measurement worthless. https://armature.tech/library/how-to-get-recommended-by-cursor https://armature.tech/library/how-to-get-recommended-by-cursor Thu, 03 Sep 2026 09:00:00 GMT Playbooks Cursor works with the project already open and indexed, which changes what it sees first. What moved its picks across 1,746 measured sessions. https://armature.tech/library/how-to-get-recommended-by-codex https://armature.tech/library/how-to-get-recommended-by-codex Thu, 03 Sep 2026 09:00:00 GMT Playbooks Codex searches the web in 53% of decision runs, which makes it the one agent where published content changes the outcome in weeks. https://armature.tech/library/how-to-get-recommended-by-claude-code https://armature.tech/library/how-to-get-recommended-by-claude-code Thu, 03 Sep 2026 09:00:00 GMT Playbooks Claude Code searches the web in 1.6% of decision runs and writes its own implementation in 19%. That changes which levers work. A specific playbook. https://armature.tech/library/how-personas-change-the-pick https://armature.tech/library/how-personas-change-the-pick Thu, 03 Sep 2026 09:00:00 GMT From the experiment We asked for the same thing as a vibe coder, a junior, a senior engineer and an enterprise team. The category leader changed in 14 of 18 categories. https://armature.tech/library/how-coding-agents-choose-tools https://armature.tech/library/how-coding-agents-choose-tools Thu, 03 Sep 2026 09:00:00 GMT Guides What happens between a developer's one-line request and a package landing in the lock file. The steps, and what changes the outcome. https://armature.tech/library/harness https://armature.tech/library/harness Thu, 03 Sep 2026 09:00:00 GMT Glossary The program around the model in a coding agent: how it reads files, which tools it has, how it plans. It decides which lever works. https://armature.tech/library/generative-engine-optimization https://armature.tech/library/generative-engine-optimization Thu, 03 Sep 2026 09:00:00 GMT Glossary The practice of making content that AI systems use and cite when they write answers. What the term means, how it is measured, and where it stops. https://armature.tech/library/generative-engine-optimization-for-developer-tools https://armature.tech/library/generative-engine-optimization-for-developer-tools Thu, 03 Sep 2026 09:00:00 GMT Guides GEO was built for brands that want mentioning in chat answers. Developer tools have a second, larger surface: the coding agent. https://armature.tech/library/evals-for-mcp-servers-and-clis https://armature.tech/library/evals-for-mcp-servers-and-clis Thu, 03 Sep 2026 09:00:00 GMT Guides Unit tests check your code. Evals check whether a real agent can still do the job through it. What to test, and why a judge is needed. https://armature.tech/library/evals-coding-agents-playbook https://armature.tech/library/evals-coding-agents-playbook Thu, 03 Sep 2026 09:00:00 GMT Playbooks Langfuse took 34% of 288 judged evals sessions. What the numbers say a vendor in this category should do.