Eden AI
Listed at https://www.edenai.co
💰 Token Pricing
| Type | Price | Note |
|---|---|---|
| Input | Pay provider's exact list price + 5.5% Eden AI platform fee per call | per million tokens |
| Output | Same: provider's exact list price, no markup. Cost returned in every API response (USD). | per million tokens |
🤖 Supported Models (1009)
✨ Pros
- ✓1,009 models aggregated across 33 providers (OpenAI, Anthropic, Google, DeepSeek, Mistral, Qwen, Z.ai, xAI, Meta, etc.)
- ✓5.5% platform fee (one of the lowest in the industry), zero markup on provider pricing — you pay list + 5.5%
- ✓OpenAI-compatible API: POST /v3/chat/completions with provider/model format
- ✓EU endpoint (api.eu.edenai.run) keeps data in Europe for GDPR / EU AI Act compliance
- ✓Universal AI endpoint (POST /v3/universal-ai) for OCR / STT / TTS / Translation / Vision / Image Gen
- ✓Public /v3/models endpoint: browse the full catalog without an API key
- ✓Self-serve plan: 10 req/sec default, upgradable to 15 req/sec on request
⚠️ Cons
- ×5.5% platform fee adds 5-10% to ultra-cheap models (e.g. DeepSeek $0.14/M input)
- ×Mainland China access requires overseas egress to main endpoint; EU endpoint targets European GDPR/AI-Act clients
- ×Provider price changes propagate automatically — no price lock-in
- ×No free monthly tier; new users must purchase credits to test (unlike OpenRouter's free models)
- ×Universal AI endpoint (non-LLM) has a different response shape from OpenAI — needs adapter code
🎯 Best For
Teams that need unified access to many AI providers (LLMs + OCR + Translation + Vision) with one key; enterprises with GDPR / EU AI Act compliance requirements; mid-size apps that don't want to build their own OpenRouter-style routing + billing layer
💰 Pricing & Plans
| Plan | Pricing Model | Platform Fee | Rate Limit | Best For |
|---|---|---|---|---|
| Self-serve AI API | Pay-as-you-go (Stripe prepaid credits) | 5.5% | 10 req/sec (upgradable to 15) | Individual devs, small teams, prototypes |
| Advanced AI Platform | Custom (invoiced) | Negotiated | Custom (higher) | Enterprises with bulk, SLA, private deployment |
🔧 API & Developer Experience
- •OpenAI-Compatible Endpoint: POST https://api.edenai.run/v3/chat/completions follows the OpenAI chat completions spec — drop-in replacement for any OpenAI SDK by changing the base URL.
- •Model Format: Use provider/model strings (e.g. openai/gpt-5.6-luna, anthropic/claude-fable-5-1, google/gemini-3.7-flash, zai/glm-5.3-flash). 1,009+ models currently catalogued.
- •Universal AI Endpoint: POST /v3/universal-ai for non-LLM tasks (OCR, STT, TTS, Translation, Vision, Image Generation, Moderation, AI Detection). Uses feature/subfeature/provider[/model] format.
- •Public Model Catalog: GET https://api.edenai.run/v3/models requires no API key — full JSON listing with context_length, capabilities, pricing, and regions for every model. Ideal for building a UI dropdown or pricing comparison.
- •EU Endpoint: Switch to https://api.eu.edenai.run to route exclusively through EU-cleared providers. Same API key works on both endpoints.
- •Streaming & Tools: SSE streaming supported on the LLM endpoint. Function calling, tool choice, vision input, parallel function calls, and response schemas work on capable models (e.g. claude-fable-5-1, gpt-5.6-luna).
- •Cost Visibility: Every API response includes the exact USD cost in a top-level `cost` field — easier cost attribution than scraping token counts.
🇪🇺 GDPR / EU AI Act Compliance & EU Endpoint
Eden AI is headquartered in France and operates a dedicated EU endpoint (https://api.eu.edenai.run) that routes requests exclusively through providers and models cleared for European processing. This is the strongest fit for workloads that must satisfy GDPR data-residency requirements or the EU AI Act's transparency and risk-management obligations. The same API key works on both the global endpoint and the EU endpoint — you only swap the host — and the discovery endpoint (/v3/models) on the EU host automatically filters to EU-available providers. The Advanced plan layers on private deployments (dedicated infrastructure for your workloads), custom billing for invoiced procurement, and a dedicated support channel with SLA — typical requirements for regulated industries (healthcare, finance, public sector in the EU). Provider prices update automatically so you always get the current rate; there is no markup on top of the underlying provider list price. The 5.5% platform fee is the only addition, applied at credit-purchase checkout.
🌐 Regional Availability & Latency
Eden AI's main endpoint (api.edenai.run) is hosted in Europe with global edge caching. Mainland China access typically requires a stable overseas egress (Hong Kong, Tokyo, or Singapore), with observed round-trip latency of 200-400 ms depending on carrier and routing path. There is no official mainland China mirror and no ICP filing, so production workloads serving China-based users should be paired with a domestic provider or aggregator (FreeModel, Volcano Ark, SiliconFlow) for the China-facing surface. For workloads with strict EU data-residency requirements, the EU endpoint (api.eu.edenai.run) ensures every byte stays within EU-cleared providers — useful for GDPR / EU AI Act compliance but adds 20-60 ms of cross-region latency compared to direct calls to OpenAI/Anthropic from US/EU clients. Outside China, latency from US East / EU West clients is comparable to direct provider calls, since Eden AI's edge terminates the request and forwards to the fastest available upstream provider.