Home/pyrite/Launch Plan: User Types & What They Care About

Launch Plan: User Types & What They Care Aboutnote

marketinglaunchuser-research
3 min read · Edit on Pyrite

User Types & What They Care About

Five user types, each with different pain points and hooks. See launch-messaging for the shared messaging foundation.

1. PKM Power Users (Obsidian, Anytype, Logseq, Notion personal)

Target wave: Wave 4 (PKM Capture Plugin). Do not actively target this audience until the capture plugin exists. The PKM crowd will be disappointed by a CLI-first platform without frictionless mobile capture. Waves 1-3 may attract some PKM-adjacent early adopters organically, but don't market to them directly until wave 4.

Profile: Individual knowledge workers who maintain personal knowledge bases. Already sold on structured notes and markdown. Pain points: search sucks at scale, no validation, tools are siloed, capturing new knowledge is high-friction.

Hook (wave 4): "Capture anything — photos, web clippings, voice notes, pasted text — and Pyrite turns it into structured, searchable, connected knowledge automatically."

Key capabilities to demo (wave 4):

  • Frictionless capture: image → OCR/vision → typed entry; URL → extract → classify → entry; voice → transcribe → entry
  • Mobile access through Claude app and web UI
  • AI auto-classification and tagging (the killer feature for PKM)
  • FTS5 + semantic search across everything you've captured
  • Knowledge graph showing connections across your captured knowledge
  • Schema validation and QA — your knowledge base stays clean automatically
  • Git-native versioning (they already use git for Obsidian vaults)
  • Markdown source of truth — no lock-in, portable

    • Objections to address:

  • "Why not just use Obsidian with plugins?" — Obsidian doesn't auto-classify, validate, or give your AI agent access to your knowledge
  • "I don't want to learn a CLI" — Wave 4 is web/mobile first. The CLI exists for power users and agents, not required for PKM
  • "Will this lock me in?" — Markdown files in git. You own everything. Take it anywhere.

    • ---

    2. Corporate KB Teams (Confluence, Notion team, SharePoint)

    Profile: Teams maintaining shared knowledge: engineering docs, runbooks, architectural decision records, onboarding guides. Pain points: docs rot, search is keyword-only, no validation, no CI/CD for knowledge.

    Hook: "Your team's knowledge base with version control, schema validation, and AI search — like upgrading from Google Docs to GitHub for your documentation."

    Key capabilities to demo:

  • Typed entries (ADRs, components, runbooks, standards)
  • QA validation — catch stale docs, broken links, missing fields automatically
  • Web UI for non-CLI users
  • Knowledge graph showing how everything connects
  • "We use Pyrite to build Pyrite" — the dogfooding story
  • Git workflow for knowledge review (PRs for doc changes)

    • Objections to address:

  • "Our team won't use a CLI"
  • "How does this integrate with our existing tools?"
  • "We already have Confluence/Notion"

    • ---

    3. Agentic Teams (AI engineers, agent runtime builders)

    Profile: Developers building autonomous agent systems (OpenClaw, custom Claude Code setups, Codex-based pipelines). Pain points: agents lose context between sessions, no structured memory, no way for agents to validate their own work.

    Hook: "Persistent, structured, validated memory for your AI agents — and they can build their own extensions for any domain."

    Key capabilities to demo:

  • Self-configuration loop (agent builds extension → tests → installs → populates KB)
  • Three-tier MCP permission model (read/write/admin)
  • CLI with `--format json` for agent consumption
  • QA validation on every write
  • Extension scaffolding (`pyrite extension init`)
  • Task coordination plugin (when shipped)

    • Objections to address:

  • "Can't I just use a vector database?"
  • "How does this compare to LangChain memory / MemGPT?"
  • "Is this production-ready?"

    • ---

    4. Python Developers (your network)

    Profile: Experienced Python devs who appreciate good architecture. May not have an immediate KB need but will recognize quality tooling and see applications.

    Hook: "A beautifully engineered knowledge platform — 15-method plugin protocol, SQLite FTS5, git-native storage, three-tier MCP. Worth studying even if you never use it."

    Key capabilities to demo:

  • Plugin protocol architecture (runtime checkable, 15 extension points)
  • Dual storage model (markdown source of truth, SQLite derived index)
  • Test suite (1780+ tests, TDD culture)
  • Extension system (pip-installable domain plugins)
  • Clean CLI with Click

    • Objections to address:

  • "What would I use this for?"
  • "Seems over-engineered for notes"

    • ---

    5. Automation Builders (n8n, Make, Zapier users)

    Profile: People who build automated workflows connecting services. Pain points: no good "knowledge" node in their pipelines, can't easily store and query structured data from workflows.

    Hook: "A knowledge base your automations can read and write — structured data in, smart search out."

    Key capabilities to demo:

  • REST API for workflow integration
  • MCP tools for AI nodes
  • CLI for shell-based automations
  • Typed entries mean consistent data structure
  • n8n workflow: RSS → Pyrite → search → alert

    • Objections to address:

  • "Can't I just use Airtable/Notion API?"
  • "Is there an official n8n node?"

    • Linked from

    Launch Plan: Content Matrix & User Acquisition Strategy Launch Plan: Channels & Content Types

    Links to

    Launch Plan: Content Matrix & User Acquisition Strategy Launch Messaging Foundation Launch Plan: Channels & Content Types