Pricing Specification

Status: Active Owner: @bilal @danny Last Updated: 2026-03-11 Supersedes: sessions/02-commercial-model/pricing-analysis.md, sessions/02-commercial-model/packaging-matrix.md

Overview

Envo operates two commercial motions, each with its own pricing model:

Motion	Sales Model	Pricing	Target Customer
Growth	Self-serve SaaS	Fixed monthly tiers (Starter / Pro / Premium)	SME landlords, small HMO operators
Acquisition	Agency outbound sales	Fixed monthly packages (Basic / Standard / Premium)	Mid-market landlords, letting agents, property managers

A third motion (Partner — enterprise, custom contracts) is planned but deferred until a signed contract requires it. See ADR-019 Commercial Motions Architecture and Commercial Tiers Scoping for the full three-motion strategy.

Part 1: Envo Growth (Self-Serve SaaS)

Tier Summary

Tier	Price	Properties	Tenants	SMS Cap
Starter	£19.99/mo	1—5	≤50	50/mo
Pro	£79.99/mo	6—20	51—200	100/mo
Premium	£349.99/mo	21+	200+	100/mo

All tiers include:

Full dashboard access
AI tenant conversations (24/7)
Issue management and routing
Document uploads and AI ingestion (RAG)
Email notifications (free)
Compliance tracking (expiry dates, alerts)

Metered Usage (Beyond Base Fee)

Channel	Rate	Included Free	Overage
SMS	£0.05/msg	50/mo (Starter), 100/mo (Pro/Premium)	£0.05/msg
WhatsApp	£0.011/msg	—	Per message
Voice	TBD	—	Per minute (future)
Email	Free	Unlimited	—

Feature Matrix (Growth Tiers)

Feature	Starter	Pro	Premium
Tenant Communication
Web chat	x	x	x
WhatsApp (shared Envo number)	x	x	x
WhatsApp (dedicated number)	—	x	x
Voice AI (shared Envo line)	—	x	x
Voice AI (dedicated / BYOAK)	—	—	x
SMS + Email notifications	x	x	x
Email as intake channel	—	x	x
Emergency detection + escalation	x	x	x
Configurable emergency keywords	—	x	x
Issue Management
AI auto-creation from conversations	x	x	x
Auto-categorisation (type + urgency)	x	x	x
Issue workflow + timeline	x	x	x
SLA tracking and alerts	—	x	x
Duplicate detection	—	x	x
Dashboard reply to tenants	—	x	x
Custom issue categories	—	—	x
Vendor Management
Manual vendor assignment	x	x	x
Tokenised vendor links (no login)	x	x	x
Vendor accept / decline / complete	x	x	x
Preferred vendor lists	—	x	x
Vendor ratings + performance	—	x	x
Vendor management portal	—	—	x
AI-recommended vendor matching	—	—	x
Property Management
Property + tenant CRUD	x	x	x
Document upload + storage	x	x	x
Compliance tracking	x	x	x
CSV import	x	x	x
Property activity log	—	x	x
Duplicate tenant detection	—	x	x
Bulk operations	—	x	x
AI + Intelligence
AI tenant conversations (Q&A + intake)	x	x	x
Document AI (extract + searchable)	x	x	x
Property document Q&A (RAG)	x	x	x
AI follow-up questions + photo requests	x	x	x
Hybrid retrieval (vector + keyword fusion)	—	x	x
AI quality test suite	—	—	x
Multilingual AI	—	—	Add-on
Reporting + Analytics
Dashboard home with live stats	x	x	x
Compliance alerts	x	x	x
Issue summary reports	—	x	x
Communication logs (regulatory)	—	x	x
CSV export	—	x	x
Advanced analytics + trends	—	—	x
Custom report builder	—	—	x
Team + Access
Single user (owner)	x	x	x
Team members (invite + roles)	—	Up to 5	Unlimited
Role-based access	—	x	x
SSO / SAML	—	—	x
Audit log	—	—	x
Branding + Customisation
Envo-branded tenant experience	x	x	x
Custom logo + colour palette	—	x	x
Custom email templates	—	x	x
Custom domain	—	—	x
Remove “Powered by Envo”	—	—	x
Custom SMS sender ID	—	—	x
Integrations
Envo shared infrastructure	x	x	x
BYOAK: SendGrid	—	x	x
BYOAK: LLM (Claude / OpenAI)	—	x	x
BYOAK: Twilio (WhatsApp + SMS)	—	—	x
BYOAK: VAPI / Retell (voice)	—	—	x
Webhook integrations	—	—	x
REST API access	—	—	x
Support
Self-service onboarding	x	x	x
AI Help Bot (in-app)	x	x	x
AI Helpline (phone)	x	x	x
Email support	x	x	x
Priority support	—	x	x
Dedicated onboarding session	—	x	x
Dedicated account manager	—	—	x

Cost + Margin Analysis

Tier	Base Fee	Typical Usage Cost	Margin	Notes
Starter	£19.99	~£3.62	82%	Acquisition tier; fixed costs amortised
Pro	£79.99	~£7.03	91%	Profit engine; where most users land
Premium	£349.99	~£20.93	94%	Enterprise margin; high profitability

Cost breakdown detail:

Infrastructure is mostly fixed (Supabase Pro £25/mo covers all orgs, CF Workers negligible per-org)
WhatsApp (£0.005/msg to us) and SMS (£0.0524/msg to us) are the main variable costs
Margins are healthy from month 1 across all tiers
Voice (future): £0.15/min cost, targeting 160% markup

Tier Upgrade Triggers

Condition	Action
Starter user >90 days, >3 properties	Prompt upgrade to Pro for WhatsApp dedicated number
Pro user hitting SMS overage 3+ months	Suggest WhatsApp (cheaper channel)
Pro user >500 WhatsApp msgs/mo	Suggest Premium for better rates
Pro user needing team access >5	Upgrade to Premium
Any user wanting custom domain / white-label	Premium

Onboarding Flow (Growth) — Implemented

Customer visits ehq.tech → signs up / logs in → lands on /onboarding
  Step 1: Organisation details (name, type)
  Step 2: Plan selection (Starter / Pro / Premium)
    → "Continue to payment" → POST /api/stripe/checkout
    → Redirect to Stripe Checkout (hosted page)
    → User pays
    → Stripe redirects to /onboarding?step=success&session_id={ID}
    → Server calls Stripe API to verify payment + read metadata
    → Creates organisation (plan, name, type all from Stripe — no client tampering)
  Step 3: Import properties (optional — Magic Import or Template)
    → Skip or complete → /dashboard

Safety net: Stripe webhook (checkout.session.completed) creates the org
if the browser redirect never completes (crash, closed tab, etc.).
Both paths are idempotent — no duplicate orgs.

See ADR-004 Billing Integration for the full architecture.

Billing Cycle

Billing date: Anniversary of signup
Charge timing: Auto-charged via Stripe
Payment method: Card on file (required)
Invoice: Sent to billing email with itemised breakdown

Free Trial

14-day trial on Pro features, no credit card required. Caps: 10 units (Starter), 25 units (Pro). Downgrades to selected tier after trial ends.

Part 2: Envo Acquisition (Agency Sales)

Overview

Acquisition is an agency-style outbound sales motion. Envo sales agents prospect, demo, and close landlords. Customers are onboarded via envo-admin and receive branded domains. Pricing is fixed monthly packages, not per-enquiry.

Tier Summary

Tier	Price	Enquiries/mo	Offers	Scripts	Reporting	Dashboard
Basic	£1,000/mo	500	1	1 basic script	Weekly report	Core access
Standard	£2,000/mo	1,500	Multi-offer	Multiple scripts	Weekly + strategy calls	CRM tools
Premium	£4,000/mo	3,000+	Full suite	Full copywriting	Full analytics + call logging	Full access

Feature Matrix (Acquisition Tiers)

Feature	Basic	Standard	Premium
Enquiries
Monthly enquiry cap	500	1,500	3,000+
Enquiry tracking + billable tagging	x	x	x
Enquiry analytics	—	x	x
Offers + Scripts
Offer configurations	1	Multiple	Full suite
Script templates	1 basic	Multiple	Full copywriting
Script A/B testing	—	—	x
Reporting
Weekly summary report	x	x	x
Strategy calls	—	x	x
Full analytics dashboard	—	—	x
Call logging + documentation	—	—	x
CRM + Retargeting
Core dashboard access	x	x	x
CRM tools	—	x	x
Retargeting campaigns	—	—	x
Support
Email support	x	x	x
Dedicated account manager	—	x	x
VA onboarding call	x	x	x

Enquiry Definition

An “enquiry” is a completed tenant interaction that results in a meaningful outcome (issue created, question answered, escalation to human). Not every message is an enquiry — it is a conversation that reaches resolution.

See Enquiry Billing Model for the full analysis of conversation lifecycle, edge cases, and the hybrid billing approach.

Billing categories:

Billable: issue_created, escalation, q_and_a
Non-billable: abandoned, spam, identity_failed

Sales Pipeline

Close CRM handles the entire Acquisition pipeline (prospect, outreach, demo, close)
On “Closed Won”, handoff to Envo ops
envo-admin provisions the organisation, configures pricing, sets up custom domain
VA books onboarding call for next day
Customer goes live on branded domain

Billing (Acquisition)

Two payment paths:

Stripe: Automated monthly billing against the fixed package price
Bank transfer + invoicing: For customers who prefer traditional invoicing. Requires invoice generation (PDF), manual payment reconciliation.

Part 3: Optional Credit System (Growth — V2)

Concept

Instead of metered per-message charges, Growth organisations can opt into a credit system that abstracts all costs into a single currency. 1 credit = £0.01.

When to use credits:

Organisations prefer predictable budgets
Heavy usage (bulk discount via topup packages)
White-label orgs (cleaner billing)

Credit Values

Action	Credits	Equivalent Cost
1 WhatsApp message	1.1	£0.011
1 SMS (overage)	5	£0.05
1 Voice minute	29	£0.29
1 Document ingestion	0.1	£0.001
100 emails	0	Free

Tier Credit Allowances

Tier	Monthly Allowance	Overage
Starter	300 credits (£3)	1 credit = £0.01
Pro	5,000 credits (£50)	1 credit = £0.01
Premium	20,000 credits (£200)	1 credit = £0.01

Topup Packages

Package	Credits	Cost	Savings vs Overage
Small	1,000	£9.00	10%
Medium	5,000	£40.00	20%
Large	10,000	£75.00	25%
Enterprise	Custom	Custom	Negotiable

Topups are one-time purchases (no auto-renew). Credits from topups roll over indefinitely.

Credit Priority

Monthly allowance consumed first (resets each billing cycle)
Purchased topups consumed second (carry over)

Part 4: Metering + Billing Infrastructure

Current State

Today the codebase tracks token usage at response time (tokensUsed returned from each LLM call) but has no persistent billing table — usage is logged but not aggregated for invoicing. Stripe integration (checkout + webhooks + subscription lifecycle) is merged and live on the magic-import branch. See ADR-004 Billing Integration for the implementation details.

Infrastructure Options

There are three complementary tools for metering AI usage. They solve different problems and can be layered.

Option A: Cloudflare AI Gateway

What it does: Proxies LLM API calls through Cloudflare, providing analytics, caching, rate limiting, and request logging. Sits between our Workers and the LLM providers.

URL: https://developers.cloudflare.com/ai-gateway/

Relevant for us because:

We already deploy on Cloudflare Workers
Provides per-request analytics (tokens, latency, cost) without custom code
Can cache identical requests (e.g., repeated RAG queries for the same document) to reduce LLM spend
Rate limiting protects against runaway costs from a single org
Logs every request with metadata (org ID, channel, model) for billing reconciliation

Limitations:

Observability only — does not handle Stripe billing integration
We would still need to aggregate usage per org and feed it into invoices
No built-in concept of “customer” or “subscription”

Verdict: Good for observability and cost control. Use it as the logging layer, but it does not replace billing logic.

Option B: Stripe AI Billing (LLM Token Metering)

What it does: Stripe’s new AI billing feature lets you route LLM calls through llm.stripe.com (or integrate via partner SDKs), automatically record token usage per Stripe Customer, and bill with a configurable markup.

Key capabilities (from Stripe announcements):

Route requests through Stripe’s LLM proxy or use partner integrations (Vercel, Cloudflare, OpenRouter, Helicone)
Pass a Stripe Customer ID on each request — usage is attributed automatically
Set a target markup (e.g., 30%) and Stripe calibrates pricing against current model rates
Dynamic updates: new subscribers get latest rates automatically
Coordinated migrations: existing subscribers update at renewal, not mid-cycle
AI and Token Meter SDKs for direct integration

Limitations:

Very new (early 2026 launch). Documentation is sparse and the product is evolving.
Requires routing all LLM traffic through Stripe’s proxy OR using a supported partner SDK
Our current multi-provider architecture (Claude → OpenAI → Kimi → GLM fallback chain) may not map cleanly to Stripe’s gateway model
Unclear how it handles non-LLM metered costs (WhatsApp, SMS, voice) — those would still need separate Stripe metered billing
Lock-in concern: tying LLM routing to a payment processor

Verdict: Watch closely, but do not adopt yet. The promise is excellent (automatic attribution + markup + invoicing), but the product is too early and our multi-provider fallback chain is a complication. Revisit when Stripe publishes stable docs and supports custom provider routing.

Option C: Build Our Own (Current Path)

What we need to build regardless:

usage_events table — log every billable action (WhatsApp msg, SMS, LLM call, voice minute)
monthly_usage aggregation — per-org monthly totals
Stripe metered billing — report usage to Stripe at billing time, generate itemised invoices
Credit ledger (V2) — optional credit system on top

This is the path we are on. The schema is designed in this spec (see Part 5). Cloudflare AI Gateway can feed into it as the LLM logging source.

Recommended Architecture

LLM Requests
  |
  v
Cloudflare AI Gateway (observability, caching, rate limiting)
  |
  v
Provider SDKs (Anthropic, OpenAI, etc.)
  |
  v
usage_events table (log tokens, cost, org_id)
  |
  v
monthly_usage aggregation (cron or webhook)
  |
  v
Stripe metered billing (invoice at billing date)

Phase 1: Ship billing with our own usage tracking (usage_events + monthly_usage + Stripe invoicing)
Phase 2: Add Cloudflare AI Gateway for LLM observability and caching
Phase 3: Evaluate Stripe AI billing when it matures; migrate if it simplifies the stack

AI SDK Migration Opportunity

Current state: The codebase uses raw provider SDKs (@anthropic-ai/sdk, openai) with a hand-rolled provider abstraction in lib/rag/providers/. This works but means we maintain our own:

Provider fallback chain
Streaming logic
Token counting
Tool-use orchestration
Retry + backoff logic

Opportunity: The Vercel AI SDK (ai package) provides a unified interface for all of this:

Unified provider API: Single interface across Anthropic, OpenAI, Google, Mistral, etc.
Built-in streaming: streamText(), streamObject() with backpressure
Tool use: First-class tool calling with type-safe schemas (Zod)
Structured output: generateObject() for typed JSON responses
Token tracking: Built-in usage reporting per call
Middleware: Hooks for logging, caching, rate limiting
Provider fallback: experimental_customProvider with fallback chains
Telemetry: OpenTelemetry integration for observability
Edge runtime compatible: Works on Cloudflare Workers

What this replaces in our codebase:

lib/rag/providers/ (entire directory) → AI SDK provider registry
lib/rag/generate.ts → generateText() / streamText()
lib/rag/intent.ts (LLM path) → generateObject() with Zod schema
lib/rag/summary.ts → generateText()
lib/tenant-engine/process.ts (tool orchestration) → AI SDK tool use
lib/rag/embed.ts → embed() / embedMany()
Custom retry/fallback logic → AI SDK middleware

What it does NOT replace:

lib/rag/chunk.ts — chunking is our own logic
lib/rag/extract.ts — file extraction is our own
lib/rag/retrieve.ts — pgvector queries stay as raw SQL
lib/rag/emergency.ts — keyword-based, no LLM
lib/rag/config.ts — our config stays

Billing benefit: AI SDK’s built-in token tracking per call maps directly to our usage_events table, making metering simpler and more reliable than parsing provider-specific response shapes.

Recommendation: This is a worthwhile migration. It reduces ~500 lines of provider abstraction code, gives us better streaming, and aligns with the ecosystem (Next.js + Vercel AI SDK is well-tested on Cloudflare Workers). Should be a separate Linear epic, not mixed with billing work.

Part 5: Database Schema

organisations (extended)

-- Existing columns plus:
tier          VARCHAR(20) DEFAULT 'STARTER'
                CHECK (tier IN ('STARTER', 'PRO', 'PREMIUM'))
monthly_charge_pence  INTEGER  -- 1999, 7999, 34999
billing_day_of_month  INTEGER
stripe_customer_id    VARCHAR(255)
billing_model         VARCHAR(20) DEFAULT 'METERED'
                        CHECK (billing_model IN ('METERED', 'CREDITS'))
monthly_credit_allowance INTEGER  -- 300, 5000, 20000 (if credits)
commercial_motion     VARCHAR(20) DEFAULT 'growth'
                        CHECK (commercial_motion IN ('growth', 'acquisition', 'partner'))

monthly_usage

CREATE TABLE monthly_usage (
    id                    UUID PRIMARY KEY,
    organisation_id       UUID REFERENCES organisations(id),
    month                 VARCHAR(7) NOT NULL,  -- '2026-03'
    whatsapp_count        INTEGER DEFAULT 0,
    whatsapp_cost_pence   INTEGER DEFAULT 0,
    sms_count             INTEGER DEFAULT 0,
    sms_overage_count     INTEGER DEFAULT 0,
    sms_overage_pence     INTEGER DEFAULT 0,
    email_count           INTEGER DEFAULT 0,
    llm_tokens_used       INTEGER DEFAULT 0,
    llm_cost_pence        INTEGER DEFAULT 0,
    created_at            TIMESTAMPTZ DEFAULT NOW(),
    updated_at            TIMESTAMPTZ DEFAULT NOW(),
    UNIQUE (organisation_id, month)
);

usage_events

CREATE TABLE usage_events (
    id                UUID PRIMARY KEY,
    organisation_id   UUID REFERENCES organisations(id),
    channel           VARCHAR(20) NOT NULL,  -- 'WHATSAPP', 'SMS', 'EMAIL', 'VOICE', 'LLM'
    count             INTEGER DEFAULT 1,
    credits_used      DECIMAL(10,2),  -- For credit system
    cost_pence        INTEGER,         -- For metered system
    metadata          JSONB,           -- model, tokens_in, tokens_out, etc.
    created_at        TIMESTAMPTZ DEFAULT NOW()
);

credit_ledger (V2)

CREATE TABLE credit_ledger (
    id                UUID PRIMARY KEY,
    organisation_id   UUID REFERENCES organisations(id),
    type              VARCHAR(20) NOT NULL
                        CHECK (type IN ('ALLOWANCE', 'PURCHASE', 'USAGE', 'ADJUSTMENT')),
    credits           DECIMAL(10,2) NOT NULL,  -- Negative for usage
    description       TEXT,
    usage_event_id    UUID REFERENCES usage_events(id),
    invoice_id        UUID REFERENCES invoices(id),
    balance_after     DECIMAL(10,2) NOT NULL,
    created_at        TIMESTAMPTZ DEFAULT NOW()
);

invoices

CREATE TABLE invoices (
    id                  UUID PRIMARY KEY,
    organisation_id     UUID REFERENCES organisations(id),
    type                VARCHAR(20) NOT NULL
                          CHECK (type IN ('SUBSCRIPTION', 'TOPUP', 'OVERAGE')),
    billing_month       VARCHAR(7),
    base_fee_pence      INTEGER,
    whatsapp_cost_pence INTEGER,
    sms_overage_pence   INTEGER,
    llm_cost_pence      INTEGER,
    topup_credits       INTEGER,
    topup_cost_pence    INTEGER,
    total_due_pence     INTEGER NOT NULL,
    stripe_invoice_id   VARCHAR(255),
    paid_at             TIMESTAMPTZ,
    due_date            DATE,
    created_at          TIMESTAMPTZ DEFAULT NOW()
);

enquiries (Acquisition billing)

CREATE TABLE enquiries (
    id                    UUID PRIMARY KEY,
    organisation_id       UUID REFERENCES organisations(id),
    conversation_id       UUID REFERENCES conversations(id),
    issue_id              UUID REFERENCES issues(id),
    channel               VARCHAR(30),
    billing_category      VARCHAR(30),  -- 'issue_created', 'q_and_a', 'escalation', 'abandoned', 'spam'
    billable              BOOLEAN DEFAULT true,
    billed_at             TIMESTAMPTZ,
    billing_period_start  DATE,
    billing_period_end    DATE,
    created_at            TIMESTAMPTZ DEFAULT NOW()
);

Part 6: Implementation Roadmap

Phase 1: Growth Billing (In Progress)

Phase 2: Acquisition Billing

envo-admin: organisation creation + package assignment (Basic/Standard/Premium)
envo-admin: enquiry tracking dashboard
enquiries table + billable tagging logic
Fixed monthly billing against package price (Stripe)
Invoice generation (PDF) for bank transfer customers
Weekly report generation (Basic+)
Strategy call scheduling (Standard+)

Phase 3: Observability + Credits

Cloudflare AI Gateway integration (LLM observability, caching)
Credit system (optional for Growth orgs)
Topup purchasing via Stripe
Credit analytics dashboard
Billing model switching UI (metered ←> credits)

Phase 4: AI SDK Migration (Separate Epic)

Replace lib/rag/providers/ with AI SDK provider registry
Migrate generate.ts to generateText() / streamText()
Migrate intent.ts to generateObject() with Zod schema
Migrate summary.ts to generateText()
Migrate tenant-engine tool orchestration to AI SDK tool use
Migrate embed.ts to AI SDK embedMany()
Wire AI SDK token tracking into usage_events

Phase 5: Stripe AI Billing (Evaluate)

Assess Stripe AI billing maturity (docs, partner support, multi-provider)
Prototype: route one provider through Stripe’s LLM proxy
Compare with our own metering — does it simplify or complicate?
Decision: adopt, defer, or reject

Part 7: Risks + Mitigations

Risk	Impact	Mitigation
Starter margin erosion on heavy HMOs	Revenue < cost for high-density properties	SMS cap + WhatsApp metering; upgrade prompts at usage thresholds
Acquisition enquiry definition disputes	Customer challenges billable count	Clear contractual definition + dispute process (see Enquiry Billing Model)
Stripe AI billing immaturity	Wasted integration effort	Watch-and-wait; build our own metering first
Credit system complexity	Engineering overhead, customer confusion	Ship as V2 opt-in; default to metered
AI SDK migration risk	Breaking changes during migration	Separate epic, comprehensive test coverage, provider-by-provider migration

Part 8: Open Questions

Question	Status	Notes
Voice pricing per minute	TBD	£0.15/min cost; target 160% markup (~£0.39/min)
Annual billing discount	TBD	20% (2 months free) standard in SaaS; defer until retention data exists
Acquisition overage pricing	TBD	What happens when Basic hits >500 enquiries? Auto-upgrade, overage fee, or hard cap?
HMO density multiplier	Deferred	Per-property pricing for now; revisit if margin erodes on high-density HMOs
Free tier	Rejected	Envo’s costs are usage-based; free tier bleeds cash. 14-day trial instead.

ADR-019 Commercial Motions Architecture — Three-motion strategy
ADR-020 Enquiry Billing Model — Enquiry definition decision
Enquiry Billing Model — Full enquiry billing analysis
Commercial Tiers Scoping — Detailed scope for Growth / Acquisition / Partner
ADR-004 Billing Integration — Stripe integration architecture
ADR-014 White-Label & BYOAK — BYOAK credential resolver design

EHQ Brain

Explorer

PRICING_SPEC

Pricing Specification

Overview

Part 1: Envo Growth (Self-Serve SaaS)

Tier Summary

Metered Usage (Beyond Base Fee)

Feature Matrix (Growth Tiers)

Cost + Margin Analysis

Tier Upgrade Triggers

Onboarding Flow (Growth) — Implemented

Billing Cycle

Free Trial

Part 2: Envo Acquisition (Agency Sales)

Overview

Tier Summary

Feature Matrix (Acquisition Tiers)

Enquiry Definition

Sales Pipeline

Billing (Acquisition)

Part 3: Optional Credit System (Growth — V2)

Concept

Credit Values

Tier Credit Allowances

Topup Packages

Credit Priority

Part 4: Metering + Billing Infrastructure

Current State

Infrastructure Options

Option A: Cloudflare AI Gateway

Option B: Stripe AI Billing (LLM Token Metering)

Option C: Build Our Own (Current Path)

Recommended Architecture

AI SDK Migration Opportunity

Part 5: Database Schema

organisations (extended)

monthly_usage

usage_events

credit_ledger (V2)

invoices

enquiries (Acquisition billing)

Part 6: Implementation Roadmap

Phase 1: Growth Billing (In Progress)

Phase 2: Acquisition Billing

Phase 3: Observability + Credits

Phase 4: AI SDK Migration (Separate Epic)

Phase 5: Stripe AI Billing (Evaluate)

Part 7: Risks + Mitigations

Part 8: Open Questions

Related Documents

Graph View

Table of Contents

Backlinks