OmniRail (@omnirailorg) header
OmniRail (@omnirailorg) profile picture

The first crypto-native, multimodal API for agents. Every text, image, video model, in one MCP command. https://teapawt.org/t.co/PyBDeVnxQH 0xa4ddf100e3e0d3a9dd897e89d025622b794f0c08

Joined September 2026
Photo by OmniRail (@omnirailorg)Photo by OmniRail (@omnirailorg)Photo by OmniRail (@omnirailorg)Photo by OmniRail (@omnirailorg)Photo by OmniRail (@omnirailorg)Photo by OmniRail (@omnirailorg)Photo by OmniRail (@omnirailorg)Photo by OmniRail (@omnirailorg)Photo by OmniRail (@omnirailorg)Photo by OmniRail (@omnirailorg)
The OmniRail platform and its token are now live. Your agent can write code. Now it can also render images, edit them, cut videos and ask any other model for a second opinion. OmniRail gives it all of them: 429 LLMs, image, video and editing models behind one key and one MCP server. It pays autonomously in USDG. omnirail.org CA (Robinhood chain) : 0xa4ddf100e3e0d3a9dd897e89d025622b794f0c08
55
16
3
125
19,957
Running several agents? See what each one costs πŸ“Š The Usage tab of your OmniRail dashboard breaks spending down per API key, per model and per day, over the last 7, 30 or 90 days. Give each agent its own key, and you see which one burns the budget, and on which model. Failed calls show up too, at no charge. Need it in a spreadsheet? Export a CSV with one line per call. omnirail.org/dashboard
1
2
9
172
Semantic search over your whole help center, indexed for 2 cents πŸ”Ž Turn each article into an embedding once. Then match a customer's question by meaning, not by keywords. 10,000 articles of about 500 tokens is 5M tokens. Perplexity Embed V1 0.6B costs $0.004 per million on OmniRail. That's 2 cents for the lot. Add a rerank model to put the best match first, then let a chat model write the answer. 35 embedding models, 11 rerankers, one key, one balance. omnirail.org/models
2
8
138
The Anthropic SDK on any model, for builders 🧩 POST /api/v1/messages takes Anthropic Messages requests, translates them into the OpenAI-style chat our gateway runs, and translates the reply back, stream events included. With the Anthropic SDK: base_url="omnirail.org/api", your OmniRail key, then messages.create(model="deepseek/deepseek-v4.1-flash", ...). Tools, images and PDFs go through. Claude ids like claude-sonnet-5-5 map to the catalog's Claude models, and the X-OmniRail-Model header says which one ran. usage.cost is what you paid, in dollars. count_tokens is free. omnirail.org/docs#anthropic
1
2
5
141
One product photo. Twenty scenes πŸ–ΌοΈ Send the photo you already have to an image-edit model, with one sentence per scene: on a marble counter, at the beach, in a gift box. On OmniRail, FLUX 3 Image Klein starts at $0.009 an image. Twenty variants from 18 cents. 61 image models take edits, on the same key as your chat calls. Like one of the shots? Hand it to one of our 29 video models and make it move. Ad tests, store listings, social posts. No photoshoot. omnirail.org/models
3
6
156
Uncensored models on OmniRail, for builders ✍️ GET /v1/models?kind=uncensored lists the curated open models. Each one carries "category": "uncensored" and its live price. They are plain chat models. Call POST /v1/chat/completions with an id like cognitivecomputations/dolphin-mistral-24b-venice-edition or nousresearch/hermes-4-405b. Streaming, keys, spend limits, scoped keys and x402 all work as usual. Over MCP: list_models with kind=uncensored, then the usual chat tool. The list is curated by hand. A model that leaves upstream drops out on its own. Our usage policy applies to every call. omnirail.org/docs#uncensored
3
7
153
Stop chunking. Send the whole thing πŸ“š 141 models on OmniRail read a million tokens or more in one call. That's a full codebase, a year of support tickets or a stack of contracts, in a single prompt. Grok 4.20 goes up to 2 million. It's cheaper than it sounds. Qwen3.7 Flash reads a million input tokens for 2.9 cents. Claude Haiku 5.5 for under 10. For a one-off question, no vector database to build and keep in sync. Ask, and the model sees everything. omnirail.org/models
2
3
9
217
When your agent swaps on Robinhood Chain through OmniRail, it gets the better of two searches ⚑ Every quote checks Uniswap's own pools. Then Uniswap plus the outside liquidity its Aggregator Hooks bring in, like Slipstream. The better route wins. The answer says where it trades (sources) and what Uniswap alone would have given (vs_uniswap_only). On launch day, 0.05 ETH β†’ PONS got 1.67% more tokens than Uniswap alone. One API call, one transaction to sign. Your agent signs with its own wallet, and funds never leave it. omnirail.org/docs/trade
1
3
6
191
Your agent can see πŸ‘€ 231 of the 358 chat models on OmniRail take images. Same call as text: put an image_url in the message. β†’ Send a screenshot, get the bug and where it is β†’ Send a receipt, get merchant, date and total as JSON β†’ Send a product photo, get alt text and tags Claude, GPT, Gemini, Qwen and more, behind one key. Start with a cheap model. Move to a frontier one only where it matters. omnirail.org/models
3
8
283
OmniRail on Android: phone sign-in without a wallet, for builders πŸ“± The phone starts a pairing and gets an 8-character code plus a random secret it never shows. Your signed-in computer types the code at /link, checks the device, its approximate location and network, and approves. The phone's next poll, with its secret, collects its own session, exactly once. We store only an HMAC of the code and a hash of the secret, and cap wrong codes per IP and per account. The app is a Trusted Web Activity over the real site, verified through assetlinks.json. API keys and x402 are unchanged. omnirail.org/docs#android
4
8
175
10,000 product reviews, sorted for about 6 cents πŸ“Š Ask a small model for JSON: a sentiment and a topic for each review. Mistral Nemo costs $0.028 per million input tokens on OmniRail. 10,000 reviews of about 200 tokens each is 2M tokens. That's about 6 cents, output included. JSON mode works on 324 of our models through the same response_format field, so your code gets clean JSON back. A few tricky ones need a smarter model? Change one string. Same key, same balance. omnirail.org/models
4
8
215
Teach your coding agent OmniRail once. It remembers in every session 🧠 The OmniRail Agent Skill is a small folder of instructions. Claude Code, Codex, Cursor and Gemini CLI read it on their own when a task needs it. After that, your agent knows how to call any model, make images, video and voice, stay inside your budget, and tell you what each call cost. In Claude Code it's two commands. For the others, a portable zip. No more pasting docs into the chat. omnirail.org/skills
4
11
219
Meeting notes for under 2 cents πŸŽ™οΈ Step 1: send the recording to a transcription model. Whisper Large V3 Turbo costs $0.0002 a minute on OmniRail. A one-hour meeting is 1.2 cents. Step 2: hand the transcript to a chat model for the summary and the action items. On Claude Haiku 5.5 that's about 0.2 cents more. Two calls, one key, one balance. Pick from 30 transcription models and 358 chat models. Swap either one by changing a string. Calls, interviews, podcasts, voice memos: all searchable text, for almost nothing. omnirail.org/models
3
1
5
222
Webhooks on OmniRail, for builders πŸ”” Pass webhook_url on POST /v1/videos, /v1/jobs or /v1/trade/submit (or in the MCP tools). When the job ends we POST one event: video.completed, job.completed, trade.completed (or .failed), with the same object GET returns. Signed with the Standard Webhooks format, so any existing verifier works. Retried for about 43 hours. Only public https addresses are called. Code to check the signature in Node and Python is in the docs: omnirail.org/docs#webhooks
2
3
7
180
Already using the OpenAI SDK? Change one line. baseURL: "omnirail.org/api/v1" That's it. Your code now reaches 358 chat models: Claude, GPT, Gemini, DeepSeek, Qwen and the rest. Switching models means changing one string. Streaming, tools, JSON output, reasoning and images work the same way. The same key also makes images, video, voice and embeddings. No new SDK. No card. You pay per call from a balance topped up in crypto. omnirail.org/docs
6
1
10
325
OmniRail shipped 38 updates in its first 5 weeks πŸš† Every one is dated on our changelog, from launch on September 30: β†’ Apps: Chat, the Playground, an Android app β†’ Agents: an Agent Skill, Claude Code on any model, x402, MCP β†’ Control: key limits, scoped keys, webhooks β†’ Models: Compare, Jev decision models, uncensored models β†’ Trade: swaps on Robinhood Chain, routed through Uniswap and its Aggregator Hooks All of it behind one key, paid in crypto. Follow the next ones, there's an RSS feed β†’ omnirail.org/changelog
4
4
10
296
Your coding agent, your desktop assistant, your custom agent: give it all of OmniRail in one line of config. The OmniRail MCP server exposes the whole catalog as tools. Your agent can list models, generate an image, render a video, transcribe audio or ask a different LLM for a second opinion, all billed to the same key, with the same spend caps. Claude writing your code and calling FLUX for the hero image. An agent that drafts a script, voices it and cuts the video. A research bot that asks three models and compares the answers. Agents stop being limited to the one model they run on. They get the whole menu. omnirail.org/docs
5
3
7
403
OmniRail Chat, for builders πŸ’¬ Every answer in /chat is a real POST /v1/chat/completions billed to your wallet: same hold, same price, same $RAIL discount, same line in your usage history as a call with your API key. So the cost under each answer is exactly what that call will cost your agent. Try a prompt on any model, switch models mid-thread to compare, then put the winning model id behind your key. omnirail.org/chat
1
1
7
396
Write one sentence. Get the whole campaign. Autopilot takes a brief and a budget and plans the job itself: the copy, the images, the edits, the video, the voice-over. Each step picks the right model from the live catalog, runs in parallel when it can, and shows its real cost as it lands. A product launch kit. A short ad with narration. A set of social visuals with matching captions. You describe the outcome, Autopilot ships the files, and it never spends past the budget you set. This is what "AI does the work" looks like when it has access to every model at once. omnirail.org/playground?mode…
1
2
7
239
Scoped API keys on OmniRail, for builders πŸ” POST /api/account/keys (and PATCH /api/account/keys/:id) take scopes: { modalities, models, max_output_price, expires_at, ips }. Models are exact ids or patterns like anthropic/*. Each keyed call is checked before the hold: out of scope β†’ 403 key_scope naming the rule, expired β†’ 401 key_expired. Autopilot jobs are planned and run inside the same scopes. Your agent can read its own fences: GET /v1/models?scoped=1 lists only what its key may call, and GET /v1/balance returns its scopes. Same rules over MCP. omnirail.org/docs#scopes
3
10
306
An agent without an account can still call OmniRail. A paid endpoint called without a key answers HTTP 402 with a PAYMENT-REQUIRED header asking for USDG on Robinhood Chain. The agent signs, retries, and gets its result. OmniRail settles on-chain and pays the gas. No key, no signup: omnirail.org/docs
1
3
8
345