Skip to content

7.4 billion tokens per month. 34 free LLM providers. 635 free model endpoints. All behind one /v1 endpoint, plus any custom OpenAI-compatible endpoint. Smart routing, automatic failover, encrypted keys. Personal experimentation only.

TypeScript

docs: add brand graphics to README (free-tier stats, comparison, features, router flow) (#625)

Tashfeenb2193b84cd66
diff --git a/README.md b/README.md
index 66d63538..57b0e940 100644
--- a/README.md
+++ b/README.md
@@ -28,6 +28,7 @@ Your router updates its own model catalog from a signed feed: new free models, q
- [Why this exists](#why-this-exists)
- [Supported providers](#supported-providers)
+- [How it compares](#how-it-compares)
- [Features](#features)
- [Not yet supported](#not-yet-supported)
- [Quick start](#quick-start)
@@ -55,6 +56,8 @@ The problem is that stacking them by hand is painful: twenty-eight different SDK
And the free-tier landscape shifts weekly: providers launch models, retire them, and change quotas without notice. FreeLLMAPI tracks all of that for you. The router pulls a signed model catalog from [freellmapi.co](https://freellmapi.co) on its own, so your install keeps up without a `git pull`. See [Premium (live catalog)](#premium-live-catalog) for how fast it keeps up.
+![The free tier, stacked — ~4B tokens of free inference per month across 28 providers](repo-assets/free-tier.png)
+
## Supported providers
<div align="center">
@@ -86,8 +89,16 @@ Plus a **custom** provider — point chat, embedding, image, or audio models at
The full, always-current list lives at **[freellmapi.co/models](https://freellmapi.co/models.html)** with per-model rate limits, context windows, and free-token budgets.
+## How it compares
+
+![Feature comparison against OpenRouter, LiteLLM, and Portkey](repo-assets/comparison.png)
+
+Based on public documentation, July 2026 — corrections welcome.
+
## Features
+![Feature overview](repo-assets/features.png)
+
- **OpenAI-compatible** — `POST /v1/chat/completions` and `GET /v1/models` work with the official OpenAI SDKs and any OpenAI-compatible client (LangChain, LlamaIndex, Continue, Hermes, etc.). Just change `base_url`.
- **Responses API** — `POST /v1/responses` (the wire format current Codex CLI versions require) is implemented as a translating shim over the same router, with full streaming events and tool calls.
- **Editor autocomplete** — `POST /v1/completions` translates legacy prompt/suffix requests into the same router, so VS Code ghost-text clients such as Continue can use FreeLLMAPI for inline suggestions.
@@ -729,6 +740,8 @@ Request volume, success rate, tokens in and out, average latency, and per-provid
## How it works
+![One request in, the best free model out — the fallback chain with live scores, cooldowns, and quota tracking](repo-assets/router-flow.png)
+
```
┌──────────────────┐ Bearer freellmapi-… ┌─────────────────────────┐
│ OpenAI SDK / │ ──────────────────────▶ │ Express proxy (:3001) │
diff --git a/repo-assets/comparison.png b/repo-assets/comparison.png
new file mode 100644
index 00000000..7676884c
Binary files /dev/null and b/repo-assets/comparison.png differ
diff --git a/repo-assets/features.png b/repo-assets/features.png
new file mode 100644
index 00000000..f066c4f9
Binary files /dev/null and b/repo-assets/features.png differ
diff --git a/repo-assets/free-tier.png b/repo-assets/free-tier.png
new file mode 100644
index 00000000..602c4b80
Binary files /dev/null and b/repo-assets/free-tier.png differ
diff --git a/repo-assets/router-flow.png b/repo-assets/router-flow.png
new file mode 100644
index 00000000..452edaea
Binary files /dev/null and b/repo-assets/router-flow.png differ