AI PM Lab

Measured conclusion: Deep-tested. L2 outcome: tools_ok.

Learn how AI products work: lessons, model cost comparisons, real experiments, PM guides.

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "AI PM Lab": {
      "url": "https://aipmlab.io/api/mcp",
      "transport": "streamable-http"
    }
  }
}

Measured tools (18)

• search_lessons read-only

  Find AI PM Lab lessons on a topic (e.g. RAG, evals, agents, prompt injection, cost). Returns the best matches with their key idea and link.

  Parameters:

  • query (required) — string — A topic or question, e.g. 'how do I stop hallucinations'

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_lesson read-only

  A lesson in full: the product decision it informs, its key idea and concepts, things to try with what to notice in each (from recorded runs), what a PM takes away and questions to ask engineers, with links to the lesson, its 3D tour and its challenge.

  Parameters:

  • slug (required) — string — The lesson's slug, from search_lessons

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• list_challenges read-only

  The challenges (games scored from real recorded runs, with leaderboards) and what each asks you to do.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• define_term read-only

  A plain-language definition of an AI product concept (tokens, temperature, RAG, embeddings, LLM-as-judge, prompt injection…), with the lessons that teach it.

  Parameters:

  • term (required) — string — The AI term to define, e.g. 'RAG', 'temperature' or 'prompt caching'

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• count_tokens read-only

  Count the tokens in a text (an estimate with OpenAI's cl100k tokenizer; other model families differ slightly).

  Parameters:

  • text (required) — string — The text to count, such as a system prompt or a typical user message

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• estimate_cost read-only

  What a model call costs per request, per day and per month at a given volume, from AI Gateway's live prices. Model ids look like 'openai/gpt-4.1-mini' or 'anthropic/claude-sonnet-4.5'.

  Parameters:

  • model (required) — string — A model id like 'openai/gpt-4.1-mini' or 'anthropic/claude-sonnet-4'
  • input_tokens (required) — integer — Input tokens per request: system prompt, retrieved documents, history and the user's message
  • output_tokens (required) — integer — Output tokens per request: the length of a typical reply
  • requests_per_day (required) — integer — Expected requests per day, for the monthly cost
  • cached_share — number — Share of input tokens served from the prompt cache, 0 to 1

  Hints: read-only yes · destructive no · idempotent yes · open-world yes

• compare_models read-only

  Compare several models side by side at a given volume: cost per request and per month, prices, context window, output limit, release date and knowledge cutoff, from AI Gateway's live catalog. Pass ids like 'openai/gpt-4.1-mini', or a provider name ('anthropic', 'openai', 'google') for its newest models, so you don't need to know which models are current.

  Parameters:

  • models (required) — array — Model ids like 'openai/gpt-4.1-mini', or provider names like 'anthropic' for that provider's newest models
  • input_tokens (required) — integer — Input tokens per request: system prompt, retrieved documents, history and the user's message
  • output_tokens (required) — integer — Output tokens per request: the length of a typical reply
  • requests_per_day (required) — integer — Expected requests per day, for the monthly cost
  • cached_share — number — Share of input tokens served from the prompt cache, 0 to 1

  Hints: read-only yes · destructive no · idempotent yes · open-world yes

• list_experiments read-only

  The real runs recorded on one of AI PM Lab's 3D pages or lessons, each described by its setup (e.g. which defences or settings were on). Replay one to see what happened.

  Parameters:

  • page (required) — string — A 3D page (injection, rag, evals, agent-loop, next-token, embeddings) or a lesson's slug (e.g. prompt-injection, model-routing, hallucination-and-grounding, human-in-the-loop)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• replay_experiment read-only

  What happened in one recorded real run (e.g. whether a prompt injection leaked data with given defences, or how many answers were made up without a facts sheet): the setup, each step, the result and the model's reply, with a link to run it on the site.

  Parameters:

  • page (required) — string — A 3D page (injection, rag, evals, agent-loop, next-token, embeddings) or a lesson's slug (e.g. prompt-injection, model-routing, hallucination-and-grounding, human-in-the-loop)
  • run (required) — integer — The run's number from list_experiments

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• quiz_me read-only

  A round from AI PM Lab's challenges, built from real recorded runs: Call It (predict what a model did), Spot the Failure (ship or flag a real output) or Which Wins? (pick the better of two setups). Returns the question without its answer; call again with the round id and the user's choice to mark it, see how players on the site called it, and get the next round.

  Parameters:

  • game — string — call-it, spot-the-failure or which-wins; any if left out
  • round — string — The id of the round the user is answering, from the last quiz_me result
  • answer — string — The user's choice for that round: its letter

  Hints: read-only yes · destructive no · idempotent no · open-world no

• review_ai_spec read-only

  Check an AI feature spec or PRD against what AI PM Lab teaches: approach, model and cost, prompt, grounding, output contract, tools, security, human oversight, evals, monitoring, fallbacks. Gaps link to lessons. Returns instructions to follow with the user's text.

  Parameters:

  • spec — string — The spec or PRD text

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• prep_planning_meeting read-only

  The questions a PM should ask engineers about a planned AI feature, with why each matters and what a good answer sounds like. Returns instructions to follow with the user's text.

  Parameters:

  • feature — string — A few sentences about the AI feature

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• design_eval_suite read-only

  Build an evaluation suite for a planned AI feature with AI PM Lab's method: test cases (typical, edge, out of scope, attacks), graders with a judge rubric, pass bars, and a JSON test set. Returns instructions to follow with the user's text.

  Parameters:

  • feature — string — What the AI feature does, who uses it, and what it must never do

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• write_acceptance_criteria read-only

  Acceptance criteria for an AI feature or user story that can actually be tested: Given/When/Then with measurable pass bars on a test set, must-pass safety criteria, latency and cost budgets, and example test cases, ready to paste into a ticket. Returns instructions to follow with the user's text.

  Parameters:

  • feature — string — The AI feature or user story, who uses it, and what it must never do

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• price_ai_feature read-only

  Work out an AI feature's unit economics and pricing: cost per request and per user (light, typical, heavy), levers like caching and routing, pricing options with margins, guardrails, and the assumptions to validate. Returns instructions to follow with the user's text.

  Parameters:

  • feature — string — What the feature does, the model it uses, expected usage, and how the product is priced today

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• plan_ai_launch read-only

  A phased roadmap from where an AI feature is today to launch and its first month live: frame, prototype and model, evals, hardening, pilot, launch, monitoring. Each phase has exit criteria and the tests and checks for this feature, backed by findings from AI PM Lab's recorded real runs (injection, grounding, evals, agent cost). Returns instructions to follow with the user's text.

  Parameters:

  • feature — string — What the AI feature does, who uses it, what it must never do, and where the work stands today

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• review_launch_readiness read-only

  A go/no-go review of an AI feature's launch plan: evals, grounding, attacks, approvals, cost, latency, monitoring, rollout and support, each marked ready, at risk or missing, with blockers and a launch-day watch list. Returns instructions to follow with the user's text.

  Parameters:

  • plan — string — The feature, what was tested, and the launch plan

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• review_ai_incident read-only

  A blameless post-incident review for an AI feature: where it failed (input, retrieval, prompt, model, tools, output checks), why evals and monitoring missed it, actions, and the new eval cases that would catch it next time. Returns instructions to follow with the user's text.

  Parameters:

  • incident — string — What the feature does, what went wrong, and what's known so far

  Hints: read-only yes · destructive no · idempotent yes · open-world no

Publisher website

Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.