Billing & Plans

Understand Praxiom AI's plans, credit system, compute boosts, trial milestones, and billing management.

Overview

Praxiom AI uses a credit-based billing system where credits map directly to the real cost of AI model usage. Every plan includes a monthly credit quota, and credits are consumed proportionally to what each operation actually costs — no flat per-run charges.

User-Scoped Billing

Subscriptions are tied to your user account, not to individual workspaces. This means:

One subscription covers all workspaces you are a member of
Credits are aggregated across all your workspaces
Upgrading your plan upgrades access everywhere

When you create a new workspace, it inherits your account-level plan and credit balance automatically.

Plans

Praxiom offers three paid plans plus a free trial tier. The pricing page displays them as cards with credits as the hero metric.

Plan	Monthly Credits	Price (Monthly)	Price (Yearly)
Trial	Milestone-based (up to 25)	Free	Free
Pro	100	$29/mo	$319/yr (1 month free)
Growth	300	$89/mo	$979/yr (1 month free)
Power	500	$149/mo	$1,639/yr (1 month free)

Toggle between monthly and yearly billing on the pricing page. Yearly billing saves you one full month.

Fetch full plan details with GET /api/billing/plans. Pass a workspace_id to see which plan is currently active.

What Each Plan Includes

Pro — For individual PMs getting started:

100 credits/month, 50 research sources, 20 documents/month
30-day chat history
Skills library, integrations, research graph
PDF and Notion export

Growth (most popular) — For teams with active research workflows:

300 credits/month, 200 research sources, unlimited documents
90-day chat history
API access (100 calls/day), GitHub integration
Up to 5 custom skills, priority execution, custom memory
Continuously optimized agent prompts

Power — For power users and autonomous research:

500 credits/month, 500 research sources, unlimited documents
1-year chat history, unlimited API calls
Multi-agent synthesis swarm, autonomous web research
Overnight research cycles (nightly experiments at 2 AM UTC)
Unlimited custom skills

Feature Gates

Higher-tier plans unlock additional capabilities. The backend enforces these via require_feature() (returns 403) and require_usage() (returns 402):

Feature	Pro	Growth	Power
`has_integrations`	Yes	Yes	Yes
`has_research_graph`	Yes	Yes	Yes
`has_skills`	Yes	Yes	Yes
`has_export_pdf`	Yes	Yes	Yes
`has_export_notion`	Yes	Yes	Yes
`has_api_access`	No	Yes	Yes
`has_custom_skills`	No	Yes (5 max)	Yes (unlimited)
`has_priority_execution`	No	Yes	Yes
`has_custom_memory`	No	Yes	Yes
`has_optimized_prompts`	No	Yes	Yes
`has_synthesis_swarm`	No	No	Yes
`has_fs_researcher`	No	No	Yes
`has_overnight_cycles`	No	No	Yes
`has_sso`	No	No	Yes
`has_audit_logs`	No	No	Yes

Usage Limits

Limit	Pro	Growth	Power
Research sources	50	200	500
Documents/month	20	Unlimited	Unlimited
Chat history	30 days	90 days	365 days
Custom skills	3	5	Unlimited
API calls/day	0	100	Unlimited
Cost target/month	$10	$25	$60
Token budget	2M	5M	10M

Compute Budget

Each plan has a monthly cost target and token budget. The feature gate warns at 80% utilisation and blocks new runs at 100%. The budget resets at the start of each billing period.

Cost-Proportional Credit System

Credits map directly to AI API cost. 1 credit = $0.08 USD of model API spend.

Credit Conversion Formula

credits = ceil(cost_usd / 0.08, nearest 0.5)

Where cost_usd is the actual Anthropic API cost of the operation (input tokens + output tokens at model-specific rates).

Model pricing used for conversion:

Model	Input (per 1M tokens)	Output (per 1M tokens)
Claude Haiku	$0.80	$4.00
Claude Sonnet	$3.00	$15.00
Claude Opus	$15.00	$75.00

Credit Floors

Run Status	Minimum Charge
Success	0.5 credits
Failed (with cost)	0.25 credits
Failed (no cost)	0 credits
Cancelled / partial	0.25 credits

Failed or cancelled runs are charged at 25–50% of their actual cost, not zero. This reflects model API usage that occurred before the failure.

Chat Message Floor

Every chat message is charged a minimum of 0.25 credits, regardless of the model or response length.

Pre-Run Estimation

Before running an agent, fetch a credit estimate:

GET /api/billing/estimate?workspace_id={id}&model={model}&effort={effort}

Response:

{
  "estimatedCredits": 2.5,
  "creditRangeLow": 1.0,
  "creditRangeHigh": 4.0,
  "remainingCredits": 45.5,
  "warning": "This may use most of your remaining credits",
  "blocked": false
}

Set blocked: true means insufficient credits — the run will be blocked before execution.

Trial Milestone System

Trial workspaces earn credits progressively through 3 milestones:

Milestone	Trigger	Credits Granted
`signup`	Account created	5.0
`first_run`	First successful agent run	5.0
`engagement`	3+ total agent runs	5.0

Total trial credits available: 15.0

Track milestone progress via GET /api/billing/milestones?workspace_id={id}. Each milestone can only be granted once per workspace.

Daily Soft Cap

To prevent a single user from consuming the entire workspace quota in one day, Praxiom calculates a per-user daily soft cap:

daily_soft_cap = (monthly_quota / active_members / 30) × 2

The 2× multiplier provides burst headroom for intensive sessions. Workspace admins can see the daily cap in the Usage section of Settings.

Credit Balance

GET /api/billing/credits?workspace_id={id}

Field	Description
`monthly_quota`	Credits included in your plan
`monthly_used`	Credits consumed this billing period
`monthly_remaining`	Remaining plan credits
`bonus_credits`	Credits from active compute boost packs
`total_available`	`monthly_quota` + all boosts ever granted
`total_used`	Monthly + boost credits consumed
`total_remaining`	Available credits before you are blocked
`is_blocked`	`true` when `total_remaining ≤ 0`

Compute Boost Packs

When you need more capacity without changing plans, purchase a one-time boost pack:

Pack	Price	Credits	Validity
Starter	$12	40 credits	30 days
Plus	$30	100 credits	30 days
Max	$65	215 credits	30 days

Boost credits are consumed after your monthly plan quota is exhausted. Packs expire after their validity window regardless of usage.

Purchase a Boost

POST /api/billing/boost/checkout

{
  "workspace_id": "your-workspace-uuid",
  "pack_type": "starter"
}

Returns a checkout_url — redirect the user to Stripe to complete payment.

View Active Boosts

GET /api/billing/boosts?workspace_id={id}

Each boost shows credits_added, credits_used, credits_remaining, expires_at, and is_active.

Credit Depletion Screen

When is_blocked becomes true, users see a Credit Depletion screen summarising what was accomplished with the credits consumed:

Count of each work type (research briefs, recommendations, documents, conversations)
Total runs and credits used
CTAs: Upgrade Plan and Buy a Credit Pack

Fetch the accomplishments data via GET /api/billing/accomplishments?workspace_id={id}.

Subscription Management

View Current Subscription

GET /api/billing/subscription?workspace_id={id}

Returns your current plan, status, period dates, and the full limits object.

Subscription statuses: active · trialing · past_due · cancelled

Upgrade or Change Plan

POST /api/billing/checkout

{
  "workspace_id": "your-workspace-uuid",
  "plan": "growth",
  "interval": "monthly",
  "success_url": "https://praxiomai.xyz/?billing=success",
  "cancel_url": "https://praxiomai.xyz/?billing=cancelled"
}

Returns a checkout_url for Stripe Checkout.

Manage Billing

POST /api/billing/portal

Generates a Stripe Customer Portal URL for managing payment methods, invoices, and cancellations.

Sync After Checkout

POST /api/billing/sync?workspace_id={id}

Call immediately after a Stripe checkout redirect to sync subscription state before webhooks fire.

Usage Dashboard

GET /api/billing/usage?workspace_id={id}

Returns a usage breakdown for the current billing period:

Usage Type	Display Name
`file_upload`	Research Sources
`agent_run`	Credits consumed
`document_draft`	Documents
`synthesis`	Daily synthesis operations
`recommendation`	Monthly recommendation generations
`chat_message`	Chat messages
`integration_sync`	Integration sync operations

Each item shows current, limit, and percentage usage.

Monthly Summary Emails

At the end of each billing period, workspace admins receive a summary email describing what was accomplished:

"5 research briefs, 8 conversations, 2 documents drafted — 142 credits used of 300"

The summary groups work by type using plain-language names (synthesis → "research briefs", full_pipeline → "deep research reports", etc.).

Stripe Webhooks

Praxiom processes the following Stripe webhook events at POST /api/webhooks/stripe:

checkout.session.completed — Provisions subscription or boost
customer.subscription.created — Activates new subscription
customer.subscription.updated — Syncs plan changes
customer.subscription.deleted — Marks subscription cancelled
invoice.payment_succeeded — Confirms payment, resets billing period
invoice.payment_failed — Sets subscription to past_due

The webhook endpoint has no JWT authentication — it is verified exclusively by Stripe's webhook signature. Never expose your STRIPE_WEBHOOK_SECRET.

What's Next

Models & Effort — how model choice affects credit cost
Missions — how multi-agent missions aggregate credits

Was this helpful?

PreviousDocuments & Editor NextPending Actions (Human-in-the-Loop)