Quickstart: How to Use OpenClaw with FPT AI Inference

Learn how to pair OpenClaw with FPT AI Inference — a high-performance, OpenAI-compatible inference API with data centers in Vietnam and Japan. 

What is FPT AI Inference?

FPT AI Inference gives you access to leading open-source models — Llama 4, Qwen3, DeepSeek, and more — through a single OpenAI-compatible API, backed by data centers in Vietnam and Japan. 

By pairing OpenClaw with FPT AI Inference, you get a fully local agent powered by frontier open-source models — your machine, your keys, your data. 

Get started in 5 minutes 

Prerequisites 

• An OpenClaw installation (docs.openclaw.ai/install) 

• An FPT AI API key — grab one at marketplace.fptcloud.com (new users get $100 free credit — including $70 for AI Inference, enough for millions of tokens to power your first OpenClaw workflows)

Step 1: Install OpenClaw 

Since FPT AI Inference is fully OpenAI-compatible, you can connect it to OpenClaw using a custom base URL. Open your OpenClaw config file.’ 

Step 2: Choose Model/auth provider 

Step 3: Enter FPT Marketplace API endpoint and Key 

Step 4: Default Model 

Environment note 

If the Gateway runs as a daemon (launchd / systemd), make sure FPT_API_KEY is available to that process. The easiest way is to keep it in ~/.openclaw/.env — OpenClaw loads this file automatically on startup. 

Security 

OpenClaw runs on your machine and calls external APIs — here’s what you need to know to stay secure. 

1. API key storage

Never hardcode your API key in config files or source code. Always use environment variables or the .env file that OpenClaw manages: 

# Good — stored in ~/.openclaw/.env (not committed to git) </span>
FPT_API_KEY=fpt_xxxxxxxxxxxx
# Bad — hardcoded in config
"apiKey": "fpt_xxxxxxxxxxxx"   // ❌ don't do this

If you suspect your key has been exposed, rotate it immediately at marketplace.fptcloud.com. 

2. Prompt injection awareness

OpenClaw is an autonomous agent — it reads web pages, files, messages, and executes actions. This makes it a target for prompt injection attacks, where malicious content in the environment tries to hijack the agent’s behavior. 

Best practices: 

• Use strong, instruction-following models (Llama 4, Qwen3) that are more resistant to injection 

• Review OpenClaw’s security best practices before deploying in production 

• Be cautious when allowing the agent to browse untrusted websites or process untrusted files 

• Prompt injection is an industry-wide unsolved problem — no model is fully immune 

3. Network security

All requests from OpenClaw to FPT AI Inference are transmitted over HTTPS/TLS. No plaintext API calls. Your prompts and responses are encrypted in transit.

Base URL: https://mkp-api.fptcloud.com
Protocol: HTTPS (TLS 1.2+)

4. Rate limiting & runaway protection

FPT AI Inference has built-in rate limiting and abuse protection. A misconfigured or looping agent won’t silently drain your credit. Consider setting a budget cap: 

{
agents: {
defaults: {
// Hard cap per session
maxTokensPerSession: 100000,
},
},
}

Monitor your usage in real-time at marketplace.fptcloud.com/my-account?tab=my-usage. 

Why FPT AI Inference + OpenClaw? 

• OpenAI-compatible — zero code changes, works out of the box 

• $100 free credit — new users get $100 Starter Plan to explore all services 

Supported models 

Full model list: marketplace.fptcloud.com/ 

Use cases 

OpenClaw powered by FPT AI Inference works great for: 

• Automating daily workflows via Telegram or WhatsApp 

• Coding assistant — write, test, and deploy code from chat 

• Research & summarization — browse and synthesize information 

• Business process automation — integrate with Slack, Teams, Discord 

Check out the OpenClaw Showcase at openclaw.ai/showcase for real-world examples. 

The bottom line 

You don’t have to choose between performance and security. FPT AI Inference gives you frontier open-source models with built-in rate limiting and HTTPS encryption. OpenClaw turns them into a full-featured agent that lives on your machine. 

Start free: marketplace.fptcloud.com/?free-trial 

GitHub PR Description 

Add FPT AI Inference as supported provider 

What this PR adds 

• Native onboarding support via openclaw onboard –auth-choice fpt-api-key 

• FPT AI Inference listed as a supported provider in docs and config 

• Quickstart guide: How to use OpenClaw with FPT AI Inference 

• Security section covering: API key storage, prompt injection, HTTPS/TLS, rate limiting 

Why FPT AI Inference 

FPT AI Inference is an OpenAI-compatible inference API with data centers in Vietnam and Japan, giving OpenClaw users access to frontier open-source models: 

• Models — Llama 4, Qwen3-32B, DeepSeek V3/R1, and more 

• $100 free Starter Plan — including $70 for AI Inference, enough for millions of tokens to power your first OpenClaw workflows 

Security notes included in doc 

This PR includes a dedicated Security section aligned with OpenClaw’s existing security-first positioning: 

• API key storage — best practices for .env usage, warn against hardcoding 

• Prompt injection — recommendation to use strong models + link to OpenClaw security docs 

• Network security — confirm HTTPS/TLS for all FPT AI Inference endpoints 

• Rate limiting — built-in abuse protection + budget cap config example 

OpenClaw has invested significantly in security (34 security commits in the last release). This doc reflects that priority and helps FPT AI Inference users adopt the same standard. 

Testing 

• openclaw onboard –auth-choice fpt-api-key ✅ 

• Model: fpt/meta-llama/Llama-4-Scout-17B-16E-Instruct ✅ 

• Model: fpt/Qwen/Qwen3-32B ✅ 

• Model: fpt/deepseek-ai/DeepSeek-V3 ✅ 

• Gateway: openclaw gateway run ✅ 

• Channels tested: Web UI, CLI ✅ 

Links 

FPT AI Inference: marketplace.fptcloud.com

Starter Plan ($100 free): marketplace.fptcloud.com/

Security best practices: docs.openclaw.ai/security