Can n8n agents use ATXP for payments?

Yes. n8n workflows that call paid APIs or services can route those calls through ATXP to get per-workflow spending limits, audit trails, and credential isolation. Use the HTTP Request node with ATXP's API to fund an agent account, then pass the agent's credentials to downstream API calls.

Does ATXP work with n8n cloud and self-hosted?

Yes. ATXP is an API — it works wherever you can make HTTP requests. Both n8n Cloud and self-hosted instances support the HTTP Request node, which is all you need to integrate ATXP.

How do I track what an n8n workflow spent?

ATXP provides a spending log for each agent account. Every API call routed through ATXP generates a receipt with timestamp, service called, and cost. Access these via the ATXP dashboard or API.

What n8n use cases work well with ATXP?

Any n8n workflow that calls a paid API — LLM inference, image generation, data enrichment, email sending — benefits from ATXP. Common patterns: AI research workflows with bounded LLM spend, content generation pipelines with per-run budgets, and data enrichment workflows where cost-per-record matters.

How to Add ATXP to n8n

n8n is one of the most widely used platforms for building agent workflows — and like most workflow automation tools, it has no native concept of a spending budget. Your API keys sit in credential storage and get used without limits until you notice the bill.

ATXP fixes that. This is how to wire it up.

What This Integration Does

When you add ATXP to an n8n workflow:

Each workflow (or set of workflows) gets its own agent account with a real credit balance
You set a spending limit — the workflow can’t exceed it
Every API call generates a receipt you can inspect
Your primary API keys stay out of the workflow entirely

The pattern is: fund an ATXP agent account → use that account’s credentials in your HTTP Request nodes → inspect spend via ATXP dashboard.

Prerequisites

An ATXP account (atxp.ai)
n8n (Cloud or self-hosted)
A workflow that makes paid API calls

Step 1: Create an Agent Account

In ATXP, an “agent account” is an isolated identity with its own credit balance. Create one per workflow or per workflow category — the granularity is up to you.

curl -X POST https://api.atxp.ai/v1/agents \
  -H "Authorization: Bearer $ATXP_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "n8n-research-workflow",
    "budget": 5.00,
    "currency": "usd"
  }'

This returns an agent ID and an agent API key. The agent key is what your n8n workflow will use — not your primary ATXP key.

Step 2: Store the Agent Key in n8n Credentials

In n8n:

Go to Settings → Credentials → Add New
Select Header Auth
Name it something like ATXP Agent — Research Workflow
Header name: Authorization, Value: Bearer [your-agent-key]

Step 3: Route API Calls Through ATXP

For each paid API call in your workflow, use the HTTP Request node with your ATXP agent credentials. ATXP supports pass-through routing for common APIs — you call the ATXP endpoint instead of the API directly.

For LLM calls (OpenAI, Anthropic, etc.):

POST https://api.atxp.ai/v1/proxy/openai/chat/completions
Authorization: Bearer [agent-key]

The request body is identical to the native API. ATXP proxies the call, records the cost, and deducts from the agent’s balance.

For arbitrary APIs, wrap the call using ATXP’s execution endpoint:

POST https://api.atxp.ai/v1/execute
Authorization: Bearer [agent-key]

{
  "service": "openai",
  "method": "chat.completions.create",
  "params": { ... }
}

Step 4: Add Budget Enforcement Logic (Optional)

ATXP will stop calls when the agent balance hits zero. If you want n8n to handle this gracefully — instead of getting a rejected call mid-workflow — add a balance check at the start of your workflow:

GET https://api.atxp.ai/v1/agents/{agent-id}/balance
Authorization: Bearer [your-primary-ATXP-key]

Use an IF node to check the balance and route to an error handler if it’s below a threshold.

Step 5: Inspect Spend

After any workflow run, check what it spent:

curl https://api.atxp.ai/v1/agents/{agent-id}/transactions \
  -H "Authorization: Bearer $ATXP_API_KEY"

Returns a log of every call: timestamp, service, cost, and status.

Common n8n + ATXP Patterns

Research workflow with bounded LLM spend: One ATXP agent account per research job. Fund it with $2. Workflow researches the topic, synthesizes a report. When the $2 is gone, it stops — no surprise bills.

Content generation pipeline: Fund an agent with credits for a batch run. Pipeline generates content until budget exhausted. Refill to run again. Cost per piece of content becomes predictable.

Data enrichment at scale: Each enrichment job (e.g., “enrich 500 contacts”) gets its own agent account. Cost-per-record is auditable. You can compare cost across different enrichment workflows.

Multi-workflow agent system: A parent workflow creates and funds child agent accounts for subworkflows, then collects spending reports. See how to build a multi-agent system for the broader pattern.

Why This Matters

n8n makes it easy to build powerful workflows. It does not make it easy to control what those workflows cost or what they have access to. Shared credentials mean a misconfigured workflow can use unlimited resources. No spending records means no debugging cost spikes.

ATXP gives each workflow a clean identity, a hard budget, and an audit trail — the three things you need to run production agent workflows with confidence.

Get started with ATXP — agent accounts, spending limits, and audit logs in one API.