MCP tools

askround mcp serves these tools over stdio. They run the command line in-process, so every refusal holds for an agent too. The schemas below are the package’s own; the server writes them for the machine and project it runs in, listing only the providers with a key, the product files it finds, and the person’s budget beside max_cost.

What each response costs an agent is measured on recorded rounds and held to a budget per tool: the brief of a whole round is about 300 tokens.

The tools

askround_ask

Ask one question to several AI assistants (OpenAI, Gemini, Anthropic, Perplexity) with web search, each several times, and read what they recommend. Returns a brief of a few hundred tokens: for each product the runs naming it and its best position per provider, its top claims and the rivals, each fact with the ids of the answers it came from ('openai#2'). Read an answer, or the passage behind a fact, with askround_answer; detail: true returns the full report instead. Refused before any call if the question names a product it tests, carries personal data, or is estimated above max_cost.

Argument	Type	What it is
question (required)	string	the question, as a person would ask it
products	array of string	products to report on, each 'Name,alias,domain'
providers	array of openai | gemini | anthropic | perplexity	only these providers (default: every one with a key)
runs	integer	runs per provider (default 3)
extract	boolean	read recommendations and claims with a model (one call per answer)
locales	array of string	e.g. ['BR:pt-BR']
search	boolean	web search (default true)
max_cost	number	refuse, before any call, a run estimated above this (USD)
detail	boolean	the full report, not the brief
fresh	boolean	ask again even if the same round was asked within the day; by default a recent one is returned at no cost
background	boolean	start it as a job and return at once (a round takes minutes); read it with askround_job

askround_answer

One answer from a round already asked, by the id a brief gives ('openai#2'): its full text and sources, or with `around` only the passage around that quote - the cheap way to check a fact.

Argument	Type	What it is
round (required)	string	the round's directory, as the brief's 'round' gives it
answer	string	the answer's id, e.g. 'openai#2' or 'openai@BR#2'
around	string	a quote: return only the passage around it
provider	string
run	integer
locale	string	when the round was asked in several locales

askround_run

Ask every question of a product file (a prompt set) and return a brief per question: the runs naming the product and its best position per provider, its claims, the rivals, each fact with its answers' ids. detail: true returns every question's full report instead.

Argument	Type	What it is
file (required)	string	the product file (TOML)
providers	array of openai | gemini | anthropic | perplexity	only these providers (default: every one with a key)
runs	integer
extract	boolean
max_cost	number	refuse, before any call, a run estimated above this (USD)
detail	boolean	every question's full report
background	boolean	start it as a job and return at once; read it with askround_job

askround_job

A round or prompt set started with background: true - running (with how many answers are in), done (with its brief or report), refused, failed, or interrupted (with the answers it finished, which are kept).

Argument	Type	What it is
job (required)	string	the id the start returned

askround_research

Ask a research question to every assistant with web search and merge the answers: the claims they make, how many providers support each, each provider's passage, and the sources they cited with primary ones (repositories, licence texts, papers, official docs) first. Support is agreement among assistants, not proof.

Argument	Type	What it is
question (required)	string
providers	array of openai | gemini | anthropic | perplexity	only these providers (default: every one with a key)
runs	integer	default 1
max_cost	number	refuse, before any call, a run estimated above this (USD)

askround_report

The report of a stored round or prompt set (default: the latest).

Argument	Type	What it is
round	string	a round's or set's directory

askround_diff

What changed between two runs of the same questions (default: the latest two): share and rank, rivals, cited pages, claims.

Argument	Type	What it is
old	string
new	string
product	string

askround_doctor

Which providers can run here: the key variable each reads (never its value), the model and the search tool. Makes a free check call to each.

No arguments.