# Plith

> Submit a task. Receive a structured, quality-scored deliverable.

Plith Rigor produces structured AI deliverables through proven methodology
frameworks with quality gates on every step. Each deliverable is
quality-scored and produced in minutes.

Built by Peaklight Inc. Web dashboard, REST API, and MCP server.

## What Plith Produces

**Competitive scans** — Porter's Five Forces, competitor positioning,
strategic recommendations. Multi-step research and analysis, not a
single-prompt summary.

**SWOT analyses** — evidence-backed strengths, weaknesses, opportunities,
threats. Each finding quality-scored.

**Code reviews** — codebase-aware via GitHub MCP connection. Reads your actual
files and produces structured findings with severity levels.

**Gap analyses** — current state vs target state comparison with severity
classification (critical, high, medium, low) and prioritized recommendations.

**Solution designs** — alternatives evaluation, tradeoff matrices,
implementation specifications, verification plans, rollback strategies.

**Bug diagnosis** — structured root-cause tracing via observe, diagnose,
scope, fix, verify methodology.

Plus 30+ additional deliverable types: implementation prompts, requirements
analysis, research briefs, incident response plans, data security
assessments, and more.

## How It Works

1. Classify — 40+ task types via 3-tier classifier (keyword, embedding, default)
2. Compose — 15+ proven methodology frameworks sequenced per task type
3. Route — adaptive model routing across 10+ LLM providers, best-fit per step
4. Gate — per-step quality scoring, automatic refinement, model escalation
5. Deliver — structured deliverable with quality score

## Product Pages

- [Homepage](https://plith.ai): Deliverable catalog, interactive demo
- [API Docs](https://plith.ai/api-docs): REST endpoints, MCP tools, patterns
- [Pricing](https://plith.ai/pricing): Credit-based billing, plan comparison
- [Help](https://plith.ai/help): Guides and FAQs
- [Demo](https://plith.ai/demo/competitive-scan): Real competitive scan example

## For Developers and AI Agents

- REST API: POST /api/rigor/plan (free) → POST /api/rigor/execute → GET /api/rigor/status/{id}
- MCP Server via Smithery: Claude Code, Cursor, Windsurf, any MCP client
- SSE streaming, webhook delivery, polling
- Interactive step-by-step mode for human-in-the-loop
- Integrations: GitHub, Notion, Linear, Jira, Confluence, Figma, Slack

## Pricing

Credit-based at 1,000 credits per USD. 20,000 free credits on web signup.
Credits never expire.
Plans: Free, Pro ($15/mo), Team ($79/mo), Enterprise ($399/mo).

## Security

Encrypted in transit and at rest. Row-level isolation. No training on
customer data. GDPR-compliant deletion. LLM policy on Team+. Data
residency on Enterprise.

## Links

- [Smithery MCP listing](https://smithery.ai/servers/plith/plith)
- [GitHub](https://github.com/NewsSpectator/plith)
- [Privacy](https://plith.ai/privacy) · [Terms](https://plith.ai/terms)