Skip to main content
From Rob, Mostly

LPRP: Your Quarterly Letter, Written the Way You Would Write It

4 min read

LPRP is an AI-native LP communication platform. It ingests your portfolio data, learns your voice from prior letters, and drafts an institutional-quality LP update—with full citation provenance and two approval gates before anything reaches a limited partner.

Built for emerging managers (Fund I to Fund III). SOC 2–ready infrastructure, tenant-isolated data.

The Problem It Addresses

LP reporting is often the worst part of the job, and most tools haven’t fixed it. Common pain points:

  • Manual assembly — 8–15 hours per letter, every quarter. Numbers copied from spreadsheets into Word; one wrong cell reference and an LP gets bad data. No audit trail.

  • No GP voice — Existing tools produce data tables and metric portals, not narratives. LPs get charts they didn’t ask for and miss the context and conviction that build the GP–LP relationship.

  • Always late — Letters often arrive 45–90 days post-quarter. Late reporting signals poor operations at a time when LPs are scrutinizing managers closely.

  • Doesn’t scale — Fund I managers may have 20 companies and no dedicated ops staff. Every hour on LP letters is an hour away from sourcing, portfolio work, and the next fundraise.

LPRP’s research (Q4 2025, n=24 fund managers) found the same refrain: the process works; it’s just done entirely by hand. LPRP is built to close that tool gap.

How It Works: Five Stages, Two Hard Gates

The pipeline runs in the background; you interact at decision points.

  1. Data ingest — Connect portfolio sources: upload updates, connect Airtable or Notion, forward founder emails, or sync cap table exports. LPRP builds a unified, provenance-tagged data layer for every company.

  2. AI extraction — Claude Haiku 3.5 parses each document into structured JSON. Every extracted value is tagged with source, date, and confidence score. Nothing enters the reporting layer without a provenance chain.

  3. Anomaly detection + Gate 1 — Claude Sonnet 4 cross-references portfolio data against your fund model and prior reporting. Variances above your thresholds surface as alerts with the specific numbers that triggered them. You review and clear before generation. Hard stop: generation is blocked until you clear Gate 1.

  4. Letter generation — Claude Opus 4 drafts the letter in your voice (calibrated from prior accepted letters) in three passes: narrative skeleton, full section drafting, then citation verification. Every quantitative claim is inline-cited to a source document.

  5. GP review + Gate 2 — Review in a tracked-changes editor. Edit inline; every claim shows its source on hover. Approve the final version and the platform sends to your LPs—addressed, formatted, and delivered. Hard stop: no one receives anything without your explicit sign-off.

Why LPRP

  • GP voice calibration — After two accepted letters, LPRP extracts a structured style brief (tone, sentence rhythm, citation format, hedge-word frequency). Later drafts read like you wrote them.

  • Full provenance chain — Every number traces to a source document (name, extraction date, confidence score). When an LP asks where a number came from, you can answer in seconds.

  • Anomaly detection before the draft — Variances are flagged before generation runs. You dismiss or confirm each alert at Gate 1; nothing sneaks through to the letter.

  • Two hard gates, no bypass — Nothing reaches LPs without explicit approval at Gate 2. The product treats your legal and reputational responsibility as non-negotiable.

  • Completeness scoring — Before generation, you see which companies are missing required metrics, which have stale data (>90 days), and which cap table entries aren’t verified. Generation is blocked on incomplete data, not silently degraded.

  • Scales from Fund I to Fund V — Built for emerging managers with 20 companies and no ops staff; grows with you to multi-fund platforms and 60+ portfolio companies.

Pricing

Simple, transparent pricing—no per-letter or per-LP fees. Plans include full data provenance, two GP review gates, anomaly detection, LP email delivery, and audit trail.

  • Starter — $200/mo ($2,400/yr): up to 15 companies, 4 quarterly letters, 1 fund, email delivery, standard support.
  • Professional — $400/mo ($4,800/yr): up to 30 companies, 2 funds, GP voice calibration after 2 letters, advanced anomaly detection, priority support.
  • Growth — $800/mo ($9,600/yr): up to 60 companies, 5 funds, unlimited letters, Carta & Pulley integration, LP portal access (Q3 2026).
  • Enterprise — Custom from $18,000/yr: unlimited companies and funds, white-label, API, SLA, dedicated account management.

Who It’s For

Emerging managers (Fund I to Fund III) who spend 8–15 hours per letter and want institutional-quality LP updates without losing their voice or control. Anyone who wants every number traceable and no LP communication without explicit approval.

Link

LPRP → https://www.lprp.online

Early access cohort onboarding in Q2 2026. Request access on the site; no credit card required, onboarding within 48 hours.