llms.txt
+15 −10
− - Catalog MCP server (discovery only): `https://avva.chat/mcp-catalog` — tools `search` and `fetch` return experts with their profile and per-expert connector URLs. Consulting runs on the expert's own connector, not here.− - Advisor MCP server: `https://avva.chat/mcp-advisor` — one connection for every public model: `search` and `fetch` find them, and the consulting tools each take the chosen model's slug as `model_slug`. Each call runs on that model's own runtime, with its limits and keys.− - Kirill Dobrinov — B2B Product Management: `https://avva.chat/mcp/b2b-product-management-kirill-dobrinov` ([page](https://avva.chat/experts/b2b-product-management-kirill-dobrinov))− - Román Tatárnikov — Product Manager (Fintech & Trading): `https://avva.chat/mcp/product-manager-fintech-trading-roman-tatarnikov` ([page](https://avva.chat/experts/product-manager-fintech-trading-roman-tatarnikov))− - Andrey Koh — B2B SaaS Product Management, Product Strategy & Analytics: `https://avva.chat/mcp/b2b-saas-product-management-product-strategy` ([page](https://avva.chat/experts/b2b-saas-product-management-product-strategy))− - [How to turn your expertise into an MCP, without writing code](https://avva.chat/blog/turn-your-expertise-into-an-mcp): Choose a field, review decisions from your work, remove private details, and publish a Decision Model your AI can connect to.− - [System prompts and past decisions: what each one carries](https://avva.chat/blog/system-prompt-vs-past-decisions): Use a worked decision to specify a tradeoff, see when a prompt is enough, and check a connected model before relying on it.− - MCP runtime: `https://avva.chat/mcp/<slug>` — packets carrying how the expert takes up a decision (where their model records it), their past decisions and what settled each, then their stop rules and criteria, for your agent to reason with rather than a finished answer (expert_questions, red_flag_check, review, compare_options, about_expert), plus `search` and `fetch` for browsing and citation. Without a key a response is a slice: the rules and, from `review` and `compare_options`, five cases − - Limits and keys: without a key, 40 requests a day, per model; with a key, 1,000 requests a day, per model. A key is issued after your user signs in on the expert's page (`https://avva.chat/experts/<slug>`) and lasts 90 days; a key a workspace member issues for an organization's model lasts 365 days. To use it, pass it as the `Authorization: Bearer <key>` header, or — where a connector dialog has no header field (ChatGPT), or the URL is simply the easier line (the Claude app) — on the end of th− - Studio extractor: `https://avva.chat/mcp-studio` — expert-side build path (not for buyers).+ - Catalog MCP server (discovery only): `https://mcp.avva.chat/mcp-catalog` — tools `search` and `fetch` return experts with their profile and per-expert connector URLs. Consulting runs on the expert's own connector, not here.+ - Advisor MCP server: `https://mcp.avva.chat/mcp-advisor` — one connection for every public model: `search` and `fetch` find them, and the consulting tools each take the chosen model's slug as `model_slug`. Each call runs on that model's own runtime, with its limits and keys.+ - Kirill Dobrinov — B2B Product Management: `https://mcp.avva.chat/mcp/b2b-product-management-kirill-dobrinov` ([page](https://avva.chat/experts/b2b-product-management-kirill-dobrinov))+ - Román Tatárnikov — Product Manager (Fintech & Trading): `https://mcp.avva.chat/mcp/product-manager-fintech-trading-roman-tatarnikov` ([page](https://avva.chat/experts/product-manager-fintech-trading-roman-tatarnikov))+ - Andreas Koch — B2B SaaS Product Management, Product Strategy & Analytics: `https://mcp.avva.chat/mcp/b2b-saas-product-management-product-strategy` ([page](https://avva.chat/experts/b2b-saas-product-management-product-strategy))+ - [“Decision model” now means two things](https://avva.chat/blog/decision-model-two-meanings): OpenAI’s Decisions API and its rivals return a fast typed answer. An avva Decision Model is one person’s published decisions. How they differ.+ - [A decision log your AI can use: what to write down](https://avva.chat/blog/decision-log-for-ai): Five fields that let an AI reason from your past decisions, a worked entry, and interview questions for recalling calls you never wrote down.+ - [Skill, system prompt or MCP: where an expert’s judgment goes](https://avva.chat/blog/skill-vs-mcp-for-expert-judgment): Five questions that decide where one person’s way of deciding should live: in a system prompt, in a skill, or behind an MCP connection.+ - [AI clone or Decision Model: two ways to put your expertise into AI](https://avva.chat/blog/ai-clone-vs-decision-model): A clone talks to your audience in your voice. A Decision Model is read by other people’s AI and never speaks as you. How to choose.+ - [What “You are a senior product manager” changes in a prompt](https://avva.chat/blog/expert-persona-prompts): Two studies found expert personas did not raise accuracy on factual questions. What the persona line does change, and what to write under it.+ - [How to turn your expertise into an MCP, without writing code](https://avva.chat/blog/turn-your-expertise-into-an-mcp): What you need, what your assistant does, the five points where it stops for you, and what other people’s AI reads once you publish.+ - [System prompts and past decisions: what each one carries](https://avva.chat/blog/system-prompt-vs-past-decisions): A rule says what to do. A past decision says which rule gave way, at what threshold, and why. When each is enough for an agent.+ - MCP runtime: `https://mcp.avva.chat/mcp/<slug>` — packets carrying how the expert takes up a decision (where their model records it), their past decisions and what settled each, then their stop rules and criteria, for your agent to reason with rather than a finished answer (expert_questions, red_flag_check, review, compare_options, about_expert), plus `search` and `fetch` for browsing and citation. Without a key a response is a slice: the rules and, from `review` and `compare_options`, five ca+ - Limits and keys: without a key, 40 requests a day, per model; with a key, 1,000 requests a day, per model. A key is issued after your user signs in on the expert's page (`https://avva.chat/experts/<slug>`) and lasts 90 days; a key a workspace member issues for an organization's model lasts 365 days. To use it, pass it as the `Authorization: Bearer <key>` header, or — where a connector dialog has no header field (ChatGPT), or the URL is simply the easier line (the Claude app) — on the end of th+ - Studio extractor: `https://mcp.avva.chat/mcp-studio` — expert-side build path (not for buyers).