AI agent hosting

Use or build AI agents for support and assistants

Arkolink provides agent hosting and a model API. Connect ready-made or custom agents to your website or app through the API, with usage billed in credits.

  • AI support
  • AI assistants
  • Troubleshooting tools
  • Usage-based billing
  • No subscription
  • Prices in USD
Your appSample data

UserAre these headphones noise-cancelling?

AgentYes, with up to 30 hours per charge.

Sample: the agent replies inside your app

What Arkolink provides

Agent hosting

Arkolink stores instructions, support greetings, example conversations, and Lorebooks so your app can use ready-made agents or your own support tools via API.

Read agent docs →
What one agent holdsSample data
Instructions
You are the support assistant for these headphones. Answer feature, spec, and post-purchase questions using the product entries; if something is not covered, ask the user to contact human support.
Support greeting
Hi, what would you like to know about these headphones?
Example conversation
User: Are these headphones noise-cancelling?
Agent: Yes, with up to 30 hours per charge.
Lorebook entries
Product introductionfeatures, specsProduct guidepairing, careReturn policyreturns, exchangesFAQbattery, noise

Model API

The model API accepts both OpenAI and Anthropic request formats for any available model. Each API key can restrict allowed models and set a monthly credits cap.

Read model API docs →
Example: call a modelSample data
$ curl "https://api.arkolink.com/v1/chat/completions" \
  -H "Authorization: Bearer $API_KEY" \
  -d '{"model": "model-a",
       "messages": [{"role": "user",
                     "content": "Hello"}]}'
200 OK
{
  "model": "model-a",
  "choices": [{ "message": { "content": "…" } }],
  "usage": { "prompt_tokens": 9, "completion_tokens": 12 }
}
Example: call a modelSample data
$ BASE=https://api.arkolink.com
$ curl $BASE/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -d '{"model": "model-a",
       "messages": [
         {"role": "user",
          "content": "Hello"}]}'
200 OK
{
  "model": "model-a",
  "choices": [{ "message": {
    "content": "…" } }],
  "usage": {
    "prompt_tokens": 9,
    "completion_tokens": 12 }
}

Four steps to get started

  1. Create an account

    Sign in with your email address to create an account. No merchant application is required.

    In the console
  2. Create an API key

    Create a personal key and select permissions such as reading and writing agents. Save the key when shown.

    In the console
  3. Create your AI agent

    Configure your agent’s instructions, support greeting, example conversations, and Lorebook entries.

    Through the API
  4. Connect to your product

    Integrate the agent into your website or app via API to serve your users with support or pre-sales Q&A.

    In your product

Pay for what you use

Agent conversations and model API calls are billed in credits based on each model’s rate and actual usage, with no subscription. Packs start at US$5 for 30,000 credits.

View plans and pricing →
  • US$530,000 credits
  • US$1061,200 credits
  • US$20124,800 credits
  • US$50318,000 credits

Data access and permissions

App access follows permissions

An authorized app can read your profile and wallet balance. Wallet entries are limited to that app; email and content access depend on the permissions you grant.

Admin actions are logged

Records include the actor, time, action and source. We retain necessary records for security and dispute resolution.

Users can revoke access

Users can revoke an app authorization. Its access tokens become invalid immediately.

Frequently asked questions

What is agent hosting?

Arkolink stores an agent’s instructions, support greeting, example conversations, and Lorebook entries so your website or app can serve users via API and update content anytime.

How do I use the model API?

Create a personal API key in the console, then send OpenAI- or Anthropic-formatted requests to any available model. Each key can limit allowed models and monthly credits.

How is usage charged?

Credits are deducted based on the selected model’s input and output rates and actual usage. Per-model rates are listed in the developer docs and the console model catalog.

Create an account to get started

Sign in to the console to create an API key and connect agents or models to your product.