Skip to main content

FAQs

Short answers, each with a link to the full page.

What can I build?​

Agents that answer questions, call tools and act in your apps. Each agent is an API as soon as you save it. Start with the Quickstart.

How do I start?​

Sign up for free, create an API key, then follow the Quickstart.

Which languages can I use?​

Python, with the SDK. From any other language, call the REST API. From an MCP client, use the MCP servers.

Can I use my own tools and data?​

Yes. Connect apps with Integrations, any MCP server with Remote MCP, or your own code as a custom tool. Give agents data through Files, a Knowledge base or Databases.

Can I change the model?​

Yes. Set the agent's llm; its tools, instructions and controls stay the same. See Choose the LLM.

How does pricing work?​

Prepaid credits, where 1 credit = 1 USD. You pay only when something runs: models and tools at their provider's price, and agent runs at that cost plus 20%. See Credits and billing.

What happens when I run out of credits?​

Runs are refused with the error code err.not_enough_balance until you add credits in Settings > Billing.

How do I keep costs down?​

Set a budget, give the agent only the tools it needs, and use a smaller LLM for simple tasks.

What is a workspace?​

Where your assets, API keys, credits and members live. You get a personal workspace and can create or join team workspaces. See Workspaces.

Where do agents run?​

In the aixplain cloud by default. On-prem is also available, including air-gapped setups. See Deployment.

How is my data protected?​

Data is encrypted in transit and at rest, and aixplain doesn't sell your data or train on it. In the app, sessions are stored so you can return to them. Via the API, sessions and traces are stored only if you enable sessions on the agent. When an agent calls an external model, that provider receives only what the model needs for that step. To keep data in your network, deploy on-prem. The aixplain cloud runs on AWS in the US (us-east-1) and is SOC 2 audited. See aixplain security for compliance details.

How do I control what agents do?​

Inspectors check inputs, steps and answers. allowed_actions limits what each tool can do, budgets cap each run, and rate limits cap each API key.

Why isn't my API key working?​

Check that the key is active, copied in full and from the right workspace, and that the workspace has credits. See API keys.

Where do I get help?​

Ask on Discord or email care@aixplain.com. For enterprise options such as on-prem deployment, dedicated capacity and higher limits, email the same address.

Next steps​

  • Quickstart: run your first agent in a few minutes.
  • Discord: ask the community and the aixplain team.