Skip to content

Migration guidesFrom Eden AI

From Eden AI, for your LLM traffic

Eden AI puts LLMs and expert models (OCR, moderation, translation and more) behind one key, billed from its credits or on your own provider keys. INFRO only covers the model side: chat, media, and the control around them. This page moves the LLM calls and says what stays.

Independent (Lyon); API v3; its own provider access on credits, plus BYOK on plans that include it

INFRO accounts are self-serve. The catalog, the docs, and the API contract are public, the gateway is deployed, and anything beyond v1 is marked Later. Plans and trial

Where things stand

Where it stands

Active and independent; the current API is v3 at https://api.edenai.run/v3. A different company, Eden AI Labs in San Francisco, is unrelated.

Whose account serves the request

By default Eden AI's provider access, paid from prepaid credits. BYOK adds your own provider keys, depending on plan, and the two can be mixed in routing.

What moves

/v3/chat/completions calls. /v3/universal-ai calls for expert models do not.

Eden AI covers far more than LLMs. Keep it for these.

  • OCR, document parsing, moderation, translation, NER and other expert models through /v3/universal-ai. INFRO does not serve them.
  • Provider choice per request by cost, speed or latency (routing.sort, :cost suffixes). INFRO has no per-request sorting.
  • Model access without your own provider accounts.
  • Its EU endpoint and zero-data-retention routing. INFRO's residency is on the roadmap.

The edits, one row each. Everything not listed here stays as it is.

 Eden AIINFRO
Base URLhttps://api.edenai.run/v3https://api.infro.io/v1
KeyAn Eden AI key.An INFRO gateway key; your provider keys become connections.
Model idopenai/gpt-latest, vertex/gemini-flash-latest, gpt-5.6-sol:cost.openai/gpt-5.6-sol, google/gemini-3.8-flash; no -latest aliases or sort suffixes.
Body fieldsfallbacks (up to 3), routing { sort, allowed_providers }, session_id.fallbacks (up to 5), routing.providers { allow, deny }, metadata.session_id.

Before · Eden AI

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.edenai.run/v3",
  apiKey: process.env.EDEN_AI_API_KEY,
});

await client.chat.completions.create({
  model: "openai/gpt-latest",
  messages: [{ role: "user", content: "Hello" }],
  // @ts-expect-error Eden AI extension
  fallbacks: ["anthropic/claude-sonnet-latest"],
});

After · INFRO

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.infro.io/v1",
  apiKey: process.env.INFRO_API_KEY,
});

await client.chat.completions.create({
  model: "openai/gpt-5.6-sol",
  messages: [{ role: "user", content: "Hello" }],
  // @ts-expect-error INFRO extension
  fallbacks: ["anthropic/claude-sonnet-5.5"],
});

Pin versions

INFRO ids name a model, not a family alias: replace gpt-latest with the model you mean. A route can hold the choice, so code sends one name you change in the console.

Two clients for a while

Keep the Eden AI client for /v3/universal-ai and point the LLM client at INFRO.

Cost per request

Both return a cost estimate. INFRO's says where it came from (usage.price_source) and is absent rather than zero when no rate is known.

Eden AI's objects against INFRO's, from Eden AI's public documentation read on 19 October 2026 and checked against what the gateway implements. One row says INFRO has nothing.

Eden AIWhat it does thereIn INFRO
Eden AI key and creditsdifferentOne key; provider usage paid from credits by default.A gateway key over your own connections; each provider bills you.
Fallbacksame shapeUp to 3 backup models per request.Up to 5 fallbacks, plus routes with failure rules.
Smart routingdifferentRank providers by cost, speed or latency; restrict; sticky sessions.Restrict with routing.providers; order on a route. No ranking by cost or speed.
Custom API keyssame shapePer-key budget, reset period, expiry, member, default tags.Gateway keys with spend limits, allowlists, environments, rotation; budgets at four more scopes.
Organization guardrailssame shapeModel access, rate limits and member budgets.Allowlists, budgets per member and key, request-rate tiers; content guardrails separately.
Monitoringsame shapeUsage and cost by key, member, feature, provider.Usage and estimated cost by project, key, model, user and tag, plus a per-request log.
Expert models (Universal AI)nothingOCR, moderation, NER, translation and more.Nothing.

Our reading of Eden AI’s published documentation, not a trial or a benchmark. If we have read something wrong, tell us and we will fix the row.

What INFRO does not do

  • Serve OCR, moderation, translation or other expert models.
  • Rank providers per request by cost, speed or latency.
  • Provide model access without your own provider accounts.
  • Pin traffic to the EU.

What does not come across

  • Usage history and credits on your Eden AI account.
  • Custom keys and guardrails as objects. Recreate them as gateway keys, budgets and allowlists.

Questions

Can I move only my LLM calls?
Yes. Point the chat client at INFRO and keep the Eden AI client for expert models.
Does `fallbacks` work the same way?
Nearly: INFRO takes up to 5 model ids, tried after every connection for model has failed.
Who pays for model usage on INFRO?
You, directly to your providers. INFRO is a subscription with 0% markup on model spend.

Move with us, or on your own

The migration program maps your Eden AI setup to INFRO connections, routes and budgets on a call and moves the first service with you. Your providers keep billing you directly; INFRO is a subscription and adds 0% to model spend.

Start 14-day trial Book a migration sessionRead the quickstart

Sources

Every statement on this page about Eden AI comes from one of these, read on 19 October 2026. Nothing is inferred from a trial, a benchmark, or a conversation.