Sherpa by Pocket FM

Sherpa by Pocket FM is a beta AI writing partner for serialized fiction. It plans arcs and characters, gives editorial feedback on drafts, localizes stories culturally and produces episodes as narrated audio dramas.

Sherpa by Pocket FM is a writing environment for serialized fiction, built around the idea that a story idea should be talked through rather than typed out from a blank page. A writer opens a project, describes a premise in a sentence or two, and Sherpa responds with follow-up questions, a sketched story world and escalating stakes. From there the tool moves through distinct phases: discussing ideas, drafting, reviewing feedback, checking the storyboard and sending the finished episode into audio production. The engine underneath is Pocket FM's Narrative World Model, which converts finished chapters into structured records of characters, relationships, locations, objects, world rules, scenes and open plot threads. Those records are wired into a temporal knowledge graph, so a question asked at chapter twelve only draws on evidence available up to that point, and the system can separate when an event happened from when it was revealed. Purpose-built agents sit on top: a Planner that keeps arcs and episode plans coherent across a long series, a Feedback Agent that scores engagement, readability, prose, coherence and pacing, and a Storyboard Agent that holds character, setting and arc details in one place as the draft changes. Sherpa is aimed at serial fiction writers rather than one-off short pieces. Episode structure, recap handling and continuity tracking all point toward authors producing ongoing series, often for audio. Two production features extend past the manuscript: a localization mode that adapts names, idioms and cultural references instead of swapping words, and an audio drama generator that turns drafts into performed narration with distinct voices and keeps one voice consistent across episodes. Writers comparing it to Sudowrite or Fabula will find the emphasis on long-form memory and audio output rather than prose polishing alone. It is in beta, and Pocket FM publishes research papers and preference comparisons on the product page to support its quality claims.