What is an MCP Elicitation?

MCP elicitation is a Model Context Protocol feature that lets a server pause mid-task and ask the connected client to collect missing information from the human user before continuing.

Without elicitation, a server that lacks a required parameter has only two bad options: guess, or fail outright. Elicitation gives it a third path — it sends a structured request back through the client describing what it needs (a date, a confirmation, a choice among options), and the client renders that as a prompt to the person at the keyboard. The server then resumes with whatever answer comes back, rather than having to encode every possible input path in advance. This matters because most real workflows aren't fully specified upfront. A travel-booking server might not know which of three flight options the user prefers until it asks; a deployment tool might need explicit human confirmation before touching production. Elicitation formalizes that back-and-forth as part of the protocol instead of leaving it to ad hoc chat parsing, which keeps the human clearly in the loop for decisions that matter. The tradeoff is latency and complexity: every elicitation round-trip pauses execution and depends on a responsive human, so well-designed servers reserve it for genuinely ambiguous or high-stakes moments rather than routine parameter-filling.

In practice with Neotask

When a Neotask agent runs a skill that needs a choice only a person can make — which invoice to reconcile, which calendar to book against — it uses MCP elicitation to surface that choice inline in the chat rather than failing the task or silently picking one. The user answers once, and the agent's tool call resumes with that answer already applied.

Related terms

Start free

Plans

Free

$0/mo

Download without a card and start for free.

Individual

$50/mo

The full personal agent platform for one person.

Enterprise

$200/mo

Multiple workspaces and capacity for larger teams.

Continue