Docket

An AI assistant for law firms that use Clio. Users chat through the web interface, Microsoft Teams, Slack, or MCP clients. The bot can answer questions about cases, look up firm procedures, and execute operations in Clio.

How It Works

User → Chat Channel → Worker → Durable Object → AI + Clio → Response

Each organization gets its own Durable Object that manages conversations, settings, and Clio credentials. The Worker routes incoming messages to the right org's DO.

Storage

Store	Scope	Contents
D1	Shared	Auth, org registry, Knowledge Base chunks
Vectorize	Shared (filtered)	Embeddings for KB and org documents
DO SQLite	Per-org	Conversations, messages, settings, Clio schema cache
DO KV	Per-org	Encrypted Clio OAuth tokens
R2	Per-org paths	Uploaded documents, audit logs

Getting Started

git clone <repo-url>
cd docket
npm install

Copy the example env files:

cp apps/api/.dev.vars.example apps/api/.dev.vars
cp apps/web/.dev.vars.example apps/web/.dev.vars

Create a .env file for the web app (Vite uses .env, Wrangler uses .dev.vars):

echo "VITE_API_URL=http://localhost:8787" > apps/web/.env

Edit apps/api/.dev.vars with your Clio credentials and secrets.

Run the API and web app in separate terminals:

npm run dev:api   # http://localhost:8787
npm run dev:web   # http://localhost:5173

Running Tests

npm test                    # Unit tests (all packages)
npm run test:e2e            # End-to-end tests
npm run test:all            # Both

# Web app specific
cd apps/web
npm test                    # Unit tests
npm run test:integration    # Integration tests (requires API running)
npm run test:e2e            # Playwright E2E tests
npm run test:e2e:ui         # Playwright with interactive UI

Integration tests require the API server running on localhost:8787.

E2E Testing with Authentication

E2E tests use Playwright. Tests requiring authentication use storage state — a saved browser session that skips login:

cd apps/web

# 1. Generate auth state (login manually, state saved to .auth/)
npx playwright test --project=setup

# 2. Run authenticated tests
npm run test:e2e

The setup project logs in once and saves cookies/localStorage to .auth/user.json. Subsequent tests load this state to skip login.

Test structure:

• test/e2e/auth-and-org.spec.ts — Signup, org creation, member invitation flows
• playwright.config.ts — Test configuration with setup project

Deployment

# Deploy API (api.docketadmin.com)
cd apps/api && wrangler deploy

# Deploy Web (docketadmin.com)
cd apps/web && npm run build && wrangler deploy --env production

Database Migrations

cd apps/api

# Local development
npx wrangler d1 migrations apply docket-db --local

# Production
npx wrangler d1 migrations apply docket-db --remote

Knowledge Base Architecture

The system has two sources of RAG context:

Shared Knowledge Base (developer-managed)

Component	Location	Purpose
Source files	apps/api/kb/*.md	Markdown content
Bundled manifest	src/services/kb-manifest.ts	Auto-generated imports
D1 table	kb_chunks	Chunk text and metadata
Vectorize	type: "kb" vectors	Semantic search

Org Context (org admin-managed)

Component	Location	Purpose
Source files	R2 via /api/org/context	Uploaded documents
D1 table	org_context_chunks	Chunk text and metadata
Vectorize	type: "org" + org_id	Per-org semantic search

Reseeding the Shared KB

When KB markdown files change, reseed to update D1 and Vectorize:

cd apps/api

# 1. Regenerate the manifest (bundles MD files into worker)
npm run kb:manifest

# 2. Deploy the worker
npx wrangler deploy

# 3. Trigger the seed (clears and rebuilds kb_chunks + vectors)
curl -X POST https://api.docketadmin.com/internal/seed-kb \
  -H "X-Seed-Secret: $SEED_SECRET"

The seed only affects shared KB data. Org context is untouched.

User Roles

• Owner — Full Clio access, can manage the organization and transfer ownership
• Admin — Full Clio access, can manage settings and invite users
• Member — Read-only Clio access, no org management

Documentation

Detailed specs are in /docs/00-specs/:

• 00-overview — Product overview
• 01-user-flows — User journeys
• 02-technical-foundation — Architecture
• 03-storage-schemas — Database schemas
• 04-auth — Authentication
• 05-channel-adapter — Teams/Slack/MCP adapters
• 06-durable-objects — Per-org state
• 07-knowledge-base — RAG implementation
• 08-workers-ai — LLM integration
• 09-clio-integration — Clio API
• 10-development-plan — Development phases
• 12-web-chat-interface — Web chat interface

Tech Stack

Cloudflare Workers, Durable Objects, D1, Vectorize, R2, Workers AI (Llama 3.1), React Router 7, TypeScript, Zod, Drizzle ORM, Better Auth.