The LLM gateway built for AI agents. One URL. Every model. Pay per request. $FBOX CA : Gt4XrcDxxn8Ltt9qcvxH7r7Ss1nsHLSBG2t3Cowxpump
Joined July 2026
- Tweets97
- Following35
- Followers240
- Likes62
Pinned Tweet
Major free tier update on Farebox
you can now use 18 frontier AI models completely free, no credits required.
Just open the Playground at farebox.fun, pick any ✦ FREE model, and start chatting. That's it.
What's free right now:
⚡ Via @GroqInc ultrafast responses: Llama 3.3 70B · Llama 3.1 8B · GPT OSS 120B · Qwen 3.6 27B · Groq Compound
🆓 Via @OpenRouter free tier: NVIDIA Nemotron Ultra 550B · Super 120B · Omni 30B · Nano 30B · Nano 12B VL · Nano 9B Poolside Laguna S & XS 2.1 OpenAI GPT OSS 20B Google Gemma 4 26B & 31B Cohere North Mini Code inclusionAI Ling 3.0 Tiny
Every response shows real-time token usage and cost even when it's $0.000. You always know exactly what you're consuming.
Try all 18 free models → farebox.fun/playground
@GroqInc @NVIDIAAI @poolsideai @OpenAI @GeminiApp @AIatMeta @TheInclusionAI
A clear breakdown of how Farebox Agents generate on-chain revenue for their creators through x402 capability payments.
When you create a Farebox Agent, you can add a revenue wallet a Solana address that receives payment every time someone calls one of your agent's capabilities.
This works through x402, a Solana-native micropayment protocol built into the Farebox gateway. Each callable skill or endpoint on your agent has a price attached to it, denominated in SOL or $FBOX. When a user or developer triggers that capability, the payment is made on-chain, instantly, and settled directly to your wallet. No invoices. No monthly payouts. No middleman holding your money.Think about what that means in practice.
You build a Pump.fun token scanner agent once. You attach the Token Scanner and Risk Analysis skills. You set it to public. You claim a .fbox slug. From that point forward, every trader who calls your agent's analysis endpoint pays you automatically on-chain, per request.
Your agent runs on Farebox infrastructure. You collect the revenue. You do nothing else.This is not passive income through staking or liquidity provision. It is revenue from a real AI service you built and deployed. The more useful your agent, the more it earns. That is the Farebox model.
How the Farebox public agent gallery works and why it changes how people discover and use AI.
When you make your Farebox Agent public, it does not just sit there. It gets listed in the agent gallery a browsable, open directory of AI agents built by real people on the Farebox network.Anyone who visits farebox.fun can discover your agent, read its description, see what skills it has, and start chatting with it immediately.
No signup wall for basic access. No friction. Just find an agent that fits what you need and use it.This creates something that did not exist before a living marketplace of specialized AI workers, each one built with a specific purpose, running on a specific model, with skills attached for structured tasks. A Pump.fun token scanner. A code reviewer. A crypto whitepaper summarizer. A business risk analyst.
All built by independent creators, all accessible from one place.For builders, this is distribution. Your agent is live, public, and discoverable the moment you create it. For users, this is choice real specialized agents instead of one generic chatbot. For the network, this is the
Build Your Own AI Agent on Farebox From Zero to Live in Minutes
Farebox Agents are fully live, and here is everything you need to know what they are, how to build one, and how to use it end to end.
A Farebox Agent is a hosted AI worker you configure once and deploy permanently on the Farebox network. It runs on whatever LLM you choose GPT-4o, Claude 3.5 Sonnet, Gemini 2.5 Flash, Llama 3, and more. You write the system prompt, define its personality and rules, attach skills, set spending caps, and it is live. No servers to manage, no infrastructure to maintain, no provider API keys to juggle. Farebox sits between your agent and every major LLM provider, handles all the routing, and bills everything unified in SOL and $FBOX.
Creating an agent takes under two minutes. Head to farebox.fun, open your dashboard, and hit Create Agent. Give it a name, write a short description for its public profile, and write your system prompt this is the core logic that defines what your agent does and how it behaves. Then pick a model from the dropdown, attach any skills you want (more on that below), set a daily and monthly spending cap so costs never run away, and choose whether to make it public so it shows up in the gallery. Finally, give it a .fbox slug a short lowercase identifier that becomes its permanent address on the network. Hit Create Agent and Mint Key, and you instantly receive a dedicated API key for that agent.
Skills are what separate a basic chatbot from a real capable agent. When you create an agent you can attach pre-built capabilities like Summarize, Draft Email, Code Review, Debug Code, Risk Analysis, Text to SQL, Sentiment Analysis, Data Extractor, Translate, and more. These are not just prompt shortcuts they are structured task modules that take defined inputs and return clean, formatted outputs. Your agent can handle complex, repeatable workflows without you writing custom logic for each one.
Using your agent is straightforward. There is a built-in chat interface at farebox.fun/agent/your-slug so anyone can open it and start talking immediately. For developers, the agent exposes a fully OpenAI-compatible API endpoint at api.farebox.fun/agent/your-s… which means if you are already using the OpenAI SDK, you change the base URL and the key, and everything works without touching the rest of your code.
The .fbox slug is your agent's permanent on-chain identity. It is claimable once, tied to your wallet, and becomes the public URL and API endpoint for your agent. If you add a revenue wallet, other users and developers who call your agent's capabilities pay you directly in SOL or $FBOX per call your agent becomes a node on the network that earns passively.
Spending caps protect you from surprise costs. Set a daily limit and a monthly limit at creation. The moment either cap is reached, the agent stops accepting requests and resets automatically on schedule. You stay in full control of what your agent spends, always.
To recap what you get the moment you create an agent: a hosted AI worker running on your chosen LLM, an OpenAI-compatible API endpoint live immediately, a system prompt with skill attachments, daily and monthly spend caps, a permanent .fbox slug on-chain, a public gallery listing, an optional revenue wallet for on-chain earnings in SOL and $FBOX, and a dedicated API key minted on creation.
Build yours now at farebox.fun.
What $FBOX actually does inside the Farebox ecosystem and why it matters beyond speculation.
Most tokens exist for one reason to be traded. $FBOX was built differently. It is the native currency of the Farebox network, and it has a direct functional role in everything that happens on the platform.When you top up credits on Farebox, you can pay in $FBOX. When someone calls a paid agent capability, the transaction settles in $FBOX or SOL directly on-chain.
When agents earn from their endpoints, revenue flows back to the owner's wallet in $FBOX. The token is not sitting on the side it is moving through every interaction on the network.This matters because demand for $FBOX is tied to actual usage. Every AI request routed through the gateway, every capability call on a public agent, every credit top-up all of it creates real, functional demand for the token. Not speculation.
Not hype. Utility.As more agents go live, more developers integrate the API, and more users discover the gallery, the network grows. And $FBOX grows with it. That is how a token with real infrastructure behind it is supposed to work.
f you are building a product that uses AI, you are probably dealing with multiple provider accounts, multiple API keys, different rate limits, inconsistent response formats, and a separate billing relationship with every model provider you use. Farebox solves all of that in one integration.The Farebox gateway is fully OpenAI-compatible.
You point your existing SDK to api.farebox.fun, swap in your Farebox API key, and you have access to GPT-4o, Claude 3.5 Sonnet, Gemini 2.5 Flash, Llama 3, and every other model on the network through a single endpoint, with unified billing in SOL and $FBOX.No individual provider accounts.
No juggling API keys per model. No separate invoices from OpenAI, Anthropic, and Google. One key, one endpoint, every model. Billing is transparent, real-time, and settled on-chain.For teams building on Solana, the fit is even tighter.
Your app already works in SOL. Your users already have Solana wallets. Your billing for AI infrastructure can now run through the same rails no fiat payment processor, no credit card dependency, no geographic restrictions. Just SOL in, AI out.That is what a Solana-native AI gateway looks like. And it is live now at farebox.fun
Build Your Own AI Agent on Farebox From Zero to Live in Minutes
Farebox Agents are fully live, and here is everything you need to know what they are, how to build one, and how to use it end to end.
A Farebox Agent is a hosted AI worker you configure once and deploy permanently on the Farebox network. It runs on whatever LLM you choose GPT-4o, Claude 3.5 Sonnet, Gemini 2.5 Flash, Llama 3, and more. You write the system prompt, define its personality and rules, attach skills, set spending caps, and it is live. No servers to manage, no infrastructure to maintain, no provider API keys to juggle. Farebox sits between your agent and every major LLM provider, handles all the routing, and bills everything unified in SOL and $FBOX.
Creating an agent takes under two minutes. Head to farebox.fun, open your dashboard, and hit Create Agent. Give it a name, write a short description for its public profile, and write your system prompt this is the core logic that defines what your agent does and how it behaves. Then pick a model from the dropdown, attach any skills you want (more on that below), set a daily and monthly spending cap so costs never run away, and choose whether to make it public so it shows up in the gallery. Finally, give it a .fbox slug a short lowercase identifier that becomes its permanent address on the network. Hit Create Agent and Mint Key, and you instantly receive a dedicated API key for that agent.
Skills are what separate a basic chatbot from a real capable agent. When you create an agent you can attach pre-built capabilities like Summarize, Draft Email, Code Review, Debug Code, Risk Analysis, Text to SQL, Sentiment Analysis, Data Extractor, Translate, and more. These are not just prompt shortcuts they are structured task modules that take defined inputs and return clean, formatted outputs. Your agent can handle complex, repeatable workflows without you writing custom logic for each one.
Using your agent is straightforward. There is a built-in chat interface at farebox.fun/agent/your-slug so anyone can open it and start talking immediately. For developers, the agent exposes a fully OpenAI-compatible API endpoint at api.farebox.fun/agent/your-s… which means if you are already using the OpenAI SDK, you change the base URL and the key, and everything works without touching the rest of your code.
The .fbox slug is your agent's permanent on-chain identity. It is claimable once, tied to your wallet, and becomes the public URL and API endpoint for your agent. If you add a revenue wallet, other users and developers who call your agent's capabilities pay you directly in SOL or $FBOX per call your agent becomes a node on the network that earns passively.
Spending caps protect you from surprise costs. Set a daily limit and a monthly limit at creation. The moment either cap is reached, the agent stops accepting requests and resets automatically on schedule. You stay in full control of what your agent spends, always.
To recap what you get the moment you create an agent: a hosted AI worker running on your chosen LLM, an OpenAI-compatible API endpoint live immediately, a system prompt with skill attachments, daily and monthly spend caps, a permanent .fbox slug on-chain, a public gallery listing, an optional revenue wallet for on-chain earnings in SOL and $FBOX, and a dedicated API key minted on creation.
Build yours now at farebox.fun.
Farebox Agents aren't chatbots. They're hosted AI workers with structured skills Code Review, Risk Analysis, Token Scanner, Data Extractor, and more exposed as callable endpoints.
Developers hit them via API the same way they'd call OpenAI. The difference is your agent runs your logic, earns on every call, and you control the spend cap from day one. $FBOX
Build Your Own AI Agent on Farebox From Zero to Live in Minutes
Farebox Agents are fully live, and here is everything you need to know what they are, how to build one, and how to use it end to end.
A Farebox Agent is a hosted AI worker you configure once and deploy permanently on the Farebox network. It runs on whatever LLM you choose GPT-4o, Claude 3.5 Sonnet, Gemini 2.5 Flash, Llama 3, and more. You write the system prompt, define its personality and rules, attach skills, set spending caps, and it is live. No servers to manage, no infrastructure to maintain, no provider API keys to juggle. Farebox sits between your agent and every major LLM provider, handles all the routing, and bills everything unified in SOL and $FBOX.
Creating an agent takes under two minutes. Head to farebox.fun, open your dashboard, and hit Create Agent. Give it a name, write a short description for its public profile, and write your system prompt this is the core logic that defines what your agent does and how it behaves. Then pick a model from the dropdown, attach any skills you want (more on that below), set a daily and monthly spending cap so costs never run away, and choose whether to make it public so it shows up in the gallery. Finally, give it a .fbox slug a short lowercase identifier that becomes its permanent address on the network. Hit Create Agent and Mint Key, and you instantly receive a dedicated API key for that agent.
Skills are what separate a basic chatbot from a real capable agent. When you create an agent you can attach pre-built capabilities like Summarize, Draft Email, Code Review, Debug Code, Risk Analysis, Text to SQL, Sentiment Analysis, Data Extractor, Translate, and more. These are not just prompt shortcuts they are structured task modules that take defined inputs and return clean, formatted outputs. Your agent can handle complex, repeatable workflows without you writing custom logic for each one.
Using your agent is straightforward. There is a built-in chat interface at farebox.fun/agent/your-slug so anyone can open it and start talking immediately. For developers, the agent exposes a fully OpenAI-compatible API endpoint at api.farebox.fun/agent/your-s… which means if you are already using the OpenAI SDK, you change the base URL and the key, and everything works without touching the rest of your code.
The .fbox slug is your agent's permanent on-chain identity. It is claimable once, tied to your wallet, and becomes the public URL and API endpoint for your agent. If you add a revenue wallet, other users and developers who call your agent's capabilities pay you directly in SOL or $FBOX per call your agent becomes a node on the network that earns passively.
Spending caps protect you from surprise costs. Set a daily limit and a monthly limit at creation. The moment either cap is reached, the agent stops accepting requests and resets automatically on schedule. You stay in full control of what your agent spends, always.
To recap what you get the moment you create an agent: a hosted AI worker running on your chosen LLM, an OpenAI-compatible API endpoint live immediately, a system prompt with skill attachments, daily and monthly spend caps, a permanent .fbox slug on-chain, a public gallery listing, an optional revenue wallet for on-chain earnings in SOL and $FBOX, and a dedicated API key minted on creation.
Build yours now at farebox.fun.
Build Your Own AI Agent on Farebox From Zero to Live in Minutes
Farebox Agents are fully live, and here is everything you need to know what they are, how to build one, and how to use it end to end.
A Farebox Agent is a hosted AI worker you configure once and deploy permanently on the Farebox network. It runs on whatever LLM you choose GPT-4o, Claude 3.5 Sonnet, Gemini 2.5 Flash, Llama 3, and more. You write the system prompt, define its personality and rules, attach skills, set spending caps, and it is live. No servers to manage, no infrastructure to maintain, no provider API keys to juggle. Farebox sits between your agent and every major LLM provider, handles all the routing, and bills everything unified in SOL and $FBOX.
Creating an agent takes under two minutes. Head to farebox.fun, open your dashboard, and hit Create Agent. Give it a name, write a short description for its public profile, and write your system prompt this is the core logic that defines what your agent does and how it behaves. Then pick a model from the dropdown, attach any skills you want (more on that below), set a daily and monthly spending cap so costs never run away, and choose whether to make it public so it shows up in the gallery. Finally, give it a .fbox slug a short lowercase identifier that becomes its permanent address on the network. Hit Create Agent and Mint Key, and you instantly receive a dedicated API key for that agent.
Skills are what separate a basic chatbot from a real capable agent. When you create an agent you can attach pre-built capabilities like Summarize, Draft Email, Code Review, Debug Code, Risk Analysis, Text to SQL, Sentiment Analysis, Data Extractor, Translate, and more. These are not just prompt shortcuts they are structured task modules that take defined inputs and return clean, formatted outputs. Your agent can handle complex, repeatable workflows without you writing custom logic for each one.
Using your agent is straightforward. There is a built-in chat interface at farebox.fun/agent/your-slug so anyone can open it and start talking immediately. For developers, the agent exposes a fully OpenAI-compatible API endpoint at api.farebox.fun/agent/your-s… which means if you are already using the OpenAI SDK, you change the base URL and the key, and everything works without touching the rest of your code.
The .fbox slug is your agent's permanent on-chain identity. It is claimable once, tied to your wallet, and becomes the public URL and API endpoint for your agent. If you add a revenue wallet, other users and developers who call your agent's capabilities pay you directly in SOL or $FBOX per call your agent becomes a node on the network that earns passively.
Spending caps protect you from surprise costs. Set a daily limit and a monthly limit at creation. The moment either cap is reached, the agent stops accepting requests and resets automatically on schedule. You stay in full control of what your agent spends, always.
To recap what you get the moment you create an agent: a hosted AI worker running on your chosen LLM, an OpenAI-compatible API endpoint live immediately, a system prompt with skill attachments, daily and monthly spend caps, a permanent .fbox slug on-chain, a public gallery listing, an optional revenue wallet for on-chain earnings in SOL and $FBOX, and a dedicated API key minted on creation.
Build yours now at farebox.fun.
MAJOR UPDATE: Agents are now LIVE on Farebox.
here's what that means and why it matters
you can now create your own AI agent directly on Farebox give it a name, a system prompt, a model, and a set of skills.
it gets its own API key, its own public profile, and its own endpoint on the network.
no infra. no servers. just deploy.
One gateway. Every major LLM. Plug it into your agent in seconds.
You don't need an OpenAI key. You don't need an Anthropic account. You don't need to manage provider credentials at all.When you build an agent on Farebox, you pick your model GPT-4o, Claude 3.5 Sonnet, Gemini 2.5 Flash, Llama 3, and more and the gateway handles the rest.
Farebox sits between your agent and every major LLM provider. One integration, all models, unified billing in SOL and $FBOX.
Switch models anytime. Scale without touching infra.
Pay only for what your agent actually uses.This is what a Solana-native AI gateway looks like.
MAJOR UPDATE: Agents are now LIVE on Farebox.
here's what that means and why it matters
you can now create your own AI agent directly on Farebox give it a name, a system prompt, a model, and a set of skills.
it gets its own API key, its own public profile, and its own endpoint on the network.
no infra. no servers. just deploy.
.fbox Slug & On-Chain Identity
Your agent has a name. And it lives on-chain.
Every agent you create on Farebox gets a .fbox slug a permanent, claimable identity on the network.Once you claim it, your agent is reachable at: api.farebox.fun/agent/your-s… can find it, pay for it, and build on top of it all without asking you for anything.It's not just a URL.
It's your agent's address on the Farebox network. Tied to your wallet. Yours to keep.Claim yours before someone else does. $FBOX
MAJOR UPDATE: Agents are now LIVE on Farebox.
here's what that means and why it matters
you can now create your own AI agent directly on Farebox give it a name, a system prompt, a model, and a set of skills.
it gets its own API key, its own public profile, and its own endpoint on the network.
no infra. no servers. just deploy.
Your agent, your budget. No surprises.
One thing that kills most AI integrations is unpredictable cost.
With Farebox Agents, you set a daily cap and a monthly cap before your agent goes live. The moment it hits the limit it stops.
No overcharges. No silent spend. No "wait, how did I burn $300?"You're not just deploying an AI. You're deploying one that knows when to stop.Full control. From day one.
$FBOX
MAJOR UPDATE: Agents are now LIVE on Farebox.
here's what that means and why it matters
you can now create your own AI agent directly on Farebox give it a name, a system prompt, a model, and a set of skills.
it gets its own API key, its own public profile, and its own endpoint on the network.
no infra. no servers. just deploy.
MAJOR UPDATE: Agents are now LIVE on Farebox.
here's what that means and why it matters
you can now create your own AI agent directly on Farebox give it a name, a system prompt, a model, and a set of skills.
it gets its own API key, its own public profile, and its own endpoint on the network.
no infra. no servers. just deploy.
to recap what you get:
✅ hosted AI agent, your model + prompt
✅ skill attachments for structured tasks
✅ daily/monthly spend caps
✅ .fbox slug permanent on-chain identity
✅ x402 capability payments in SOL/$FBOX
✅ public gallery listing
✅ API key minted on creation
agents are live now at farebox.fun
create one, claim your slug, set your revenue wallet.
this is the first step toward a network of specialized agents each one ownable, callable, and earning.
more soon.
We just submitted a PR to @OOBEonSol 🤝
Farebox is now proposed as an official MCP preset in the Synapse Client SDK awaiting maintainer review.
If merged, developers using Synapse can connect to 36+ frontier LLMs (GPT-4o, Claude, Gemini, Llama, DeepSeek, and more) and pay per token in USDC on Solana, directly from their agent stack.
PR #52: github.com/OOBE-PROTOCOL/syn…
FAREBOX UPDATE $FBOX Integration Goes Live
$FBOX is no longer a roadmap item. It's enforced on every request, right now.
Here's what shipped:
Tier system fully live Hold $FBOX → your tier is read on-chain every balance check and written to the gateway automatically.
5 tiers, real enforcement on every API call:
→ Basic (0): 60 RPM baseline
→ Explorer (100K): 120 RPM · −5% off · +2% cashback
→ Builder (1M): 240 RPM · −10% off · +4% cashback
→ Operator (10M): 600 RPM · −15% off · +6% cashback
→ Architect (30M): ∞ RPM · −20% off · +8% cashback
Cashback is a real ledger entry. Every successful request triggers a credit back to your balance tagged "$FBOX cashback · [Tier]" applied on top of your discount.
Node earnings
→ $FBOX. Routing margin is now converted to $FBOX at the live Jupiter market price. Refreshes every 30s on your node dashboard.
Deposit credits with $FBOX. Send $FBOX to your deposit address
→ verified on-chain via Helius
→ converted to USD at Jupiter price
→ balance credited instantly. No manual steps. You buy $FBOX on any DEX we just accept the deposit.
UI updated across the board:
· Credits page: USDC / $FBOX tab switcher with live price preview
· Tier panel: shows your real discount %, cashback %, RPM not roadmap text
· Node dashboard: earnings displayed in $FBOX with live price
· Compute page: all references to "USDC settlement coming soon" removed
For API builders running high inference volume connect your wallet, your tier activates on the next request.
We're working on smarter earning inside Farebox.
The next step:
better visibility into what you're earning, when, and why so node operators can make informed decisions about their compute allocation. Real numbers, real performance data, no guesswork.
More compute, better routing, optimized rewards. The infrastructure is there. The tooling to maximize it is coming.
Stay close $FBOX
Farebox Performance & Reliability What We're Optimizing
Speed and trust are the two things a gateway can never compromise on. Here's what we're actively improving under the hood.
Latency:
→ On-chain SNS lookups are getting a caching layer domain status will load instantly instead of hitting the RPC on every request
→ Billing metering is async and off the hot path target added latency p95 under 50ms
→ Model routing table moves to edge cache no DB hit for every key auth
Reliability:
→ Webhook attack history will persist across server restarts no more losing audit logs on redeploy
→ Rate-limit detection extending to OpenRouter and Anthropic models every restricted model now surfaces a clear, friendly error instead of a silent failure
→ Auto-retry on rate-limit hits agents won't need to resend requests manually when a provider throttles
Transparency:
→ Per-agent analytics coming with Agent Builder cost, latency, and token usage broken down per deployed agent
→ All billing events are append-only and exportable no black boxes
Infrastructure should be boring. Ours will be. $FBOX