diff --git a/README.md b/README.md
index cfda65f..b973195 100644
--- a/README.md
+++ b/README.md
@@ -2,7 +2,7 @@
 
 Step-by-step guides for connecting [Eden AI](https://www.edenai.co/) to the tools, agents, and platforms you already use.
 
-Each guide walks through installation, configuration, and model selection so you can route your workflow through **500+ LLMs and AI models** with a single API key — with smart routing, automatic fallbacks, unified billing, and per-request cost visibility.
+Each guide walks through installation, configuration, and model selection so you can route your workflow through **700+ LLMs and AI models** with a single API key — with smart routing, automatic fallbacks, unified billing, and per-request cost visibility.
 
 > Don't have a key yet? Sign up at [app.edenai.run](https://app.edenai.run/) → **API Keys**.
 
@@ -16,6 +16,23 @@ Each guide walks through installation, configuration, and model selection so you
 | OpenCode | [opencode.md](./opencode.md) | [YouTube](https://youtu.be/OYHRY5rblz0) | [edenai.co/integrations/opencode](https://www.edenai.co/integrations/opencode) |
 | Cline | [cline.md](./cline.md) | [YouTube](https://youtu.be/_M8PwHSwgMQ) | [edenai.co/integrations/cline](https://www.edenai.co/integrations/cline) |
 
+## Coding agents (terminal & IDE)
+
+Connect any of the leading open-source coding agents to Eden AI. All use the OpenAI-compatible endpoint `https://api.edenai.run/v3` with `provider/model` naming.
+
+| Tool | Guide | Video | Mechanism |
+|---|---|---|---|
+| OpenCode | [opencode.md](./opencode.md) | [YouTube](https://youtu.be/OYHRY5rblz0) | Custom provider (`opencode.json`) |
+| Codex CLI | [codex-cli.md](./codex-cli.md) | [YouTube](https://youtu.be/sVLPbencnnM) | `config.toml` provider (`wire_api = "responses"`) |
+| OpenHands | [openhands.md](./openhands.md) | — | LiteLLM (OpenAI-compatible) |
+| Cline | [cline.md](./cline.md) | [YouTube](https://youtu.be/_M8PwHSwgMQ) | Built-in Eden AI provider |
+| Pi | [pi.md](./pi.md) | — | Custom provider (`~/.pi/agent/models.json`) |
+| Goose | [goose.md](./goose.md) | — | Custom provider / env |
+| Aider | [aider.md](./aider.md) | — | OpenAI-compatible / LiteLLM |
+| Continue | [continue.md](./continue.md) | — | Custom provider (`config.yaml`) |
+| Qwen Code | [qwen-code.md](./qwen-code.md) | — | Env (`OPENAI_BASE_URL`) |
+| Crush | [crush.md](./crush.md) | — | Custom provider (`crush.json`) |
+
 More integrations are added regularly. Open an issue if there's a tool you'd like us to cover.
 
 ---
diff --git a/aider.md b/aider.md
new file mode 100644
index 0000000..680f862
--- /dev/null
+++ b/aider.md
@@ -0,0 +1,95 @@
+# Aider + Eden AI
+
+Connect [Aider](https://aider.chat) to Eden AI and route the original terminal AI pair programmer through **700+ models** with smart routing, automatic fallbacks, and a single API key.
+
+> Get your Eden AI key at [app.edenai.run](https://app.edenai.run/) → **API Keys**.
+
+---
+
+## 1. System dependencies
+
+On a fresh Linux machine (Ubuntu / Debian / WSL):
+
+```bash
+sudo apt update
+sudo apt install -y curl git python3 python3-venv
+```
+
+## 2. Install Aider
+
+```bash
+# Recommended — isolated install via uv
+curl -LsSf https://astral.sh/uv/install.sh | sh
+uv tool install aider-chat
+
+# Or via pip
+python3 -m pip install aider-install && aider-install
+
+# Verify
+aider --version
+```
+
+## 3. Point Aider at Eden AI
+
+Aider connects to any OpenAI-compatible endpoint. Export the base URL and your key:
+
+```bash
+export OPENAI_API_BASE="https://api.edenai.run/v3"
+export OPENAI_API_KEY="YOUR_EDEN_AI_API_KEY"
+```
+
+## 4. Run Aider
+
+Aider runs on top of LiteLLM, so prefix the Eden AI model with `openai/` (this selects the OpenAI-compatible driver; it is stripped before the request):
+
+```bash
+# Fast and cheap
+aider --model openai/openai/gpt-4o-mini
+
+# Claude via Eden AI
+aider --model openai/anthropic/claude-haiku-4-5
+```
+
+Every edit now flows through Eden AI — with built-in fallbacks, unified billing, and per-request cost visibility.
+
+---
+
+## Using specific models
+
+The model string is `openai/<eden-provider>/<eden-model>`:
+
+| Aider `--model` value | Best for |
+|---|---|
+| `openai/openai/gpt-4o-mini` | Fast and cheap |
+| `openai/anthropic/claude-sonnet-4-5` | Balanced reasoning |
+| `openai/anthropic/claude-opus-4-7` | Maximum capability |
+| `openai/openai/gpt-5` | OpenAI's frontier model |
+| `openai/google/gemini-2.5-pro` | Long context, multimodal |
+
+> **Note.** Use a concrete `provider/model` with Aider. The smart-router alias `@edenai` is not recommended here — LiteLLM rejects some routed reasoning-model parameters.
+
+## Troubleshooting
+
+### `401 Unauthorized`
+Verify your Eden AI key — grab a fresh one at [app.edenai.run → API Keys](https://app.edenai.run/). Confirm `OPENAI_API_KEY` and `OPENAI_API_BASE` are exported in the same shell.
+
+### `NotFoundError` / `model not found`
+Keep the double prefix: `openai/` + the Eden AI `provider/model` (e.g. `openai/anthropic/claude-haiku-4-5`).
+
+### `BadRequestError: Unsupported value`
+You're likely targeting a reasoning model or the `@edenai` router. Switch to a concrete chat model such as `openai/openai/gpt-4o-mini`.
+
+## Uninstall
+
+```bash
+uv tool uninstall aider-chat   # or: pip uninstall aider-chat
+```
+
+---
+
+## Next steps
+
+- [Smart routing](https://www.edenai.co/docs/v3/llms/smart-routing)
+- [Fallbacks](https://www.edenai.co/docs/v3/llms/fallback)
+- [Eden AI pricing](https://www.edenai.co/pricing)
+- [Aider docs](https://aider.chat/docs/)
diff --git a/codex-cli.md b/codex-cli.md
new file mode 100644
index 0000000..9f0fcdd
--- /dev/null
+++ b/codex-cli.md
@@ -0,0 +1,118 @@
+# Codex CLI + Eden AI
+
+Connect [OpenAI Codex CLI](https://github.com/openai/codex) to Eden AI and route your terminal agent through **700+ models** with smart routing, automatic fallbacks, and a single API key.
+
+> Get your Eden AI key at [app.edenai.run](https://app.edenai.run/) → **API Keys**.
+
+Follow this tuto in the same time with our video tuto on [Youtube](https://youtu.be/sVLPbencnnM)
+
+---
+
+## 1. System dependencies
+
+On a fresh Linux machine (Ubuntu / Debian / WSL):
+
+```bash
+sudo apt update
+sudo apt install -y curl git
+```
+
+## 2. Install Codex CLI
+
+```bash
+# Option A — via npm
+npm install -g @openai/codex
+
+# Option B — via Homebrew (macOS / Linuxbrew)
+brew install codex
+
+# Verify
+codex --version
+```
+
+## 3. Add Eden AI as a custom provider
+
+Codex stores its config at `~/.codex/config.toml`. Add Eden AI as a model provider:
+
+```bash
+mkdir -p ~/.codex
+nano ~/.codex/config.toml
+```
+
+```toml
+model = "openai/gpt-4o-mini"
+model_provider = "edenai"
+
+[model_providers.edenai]
+name = "Eden AI"
+base_url = "https://api.edenai.run/v3"
+env_key = "EDENAI_API_KEY"
+wire_api = "responses"
+```
+
+Export your key so Codex can read it:
+
+```bash
+export EDENAI_API_KEY="YOUR_EDEN_AI_API_KEY"
+```
+
+> **`wire_api = "responses"` is required.** Recent Codex CLI (≥ 0.122) removed support for `wire_api = "chat"`. Eden AI implements the OpenAI **Responses API** (`/v3/responses`), so Codex connects directly — no proxy needed. The provider ids `openai`, `ollama`, `lmstudio` are reserved, so keep the id `edenai`.
+
+## 4. Run Codex
+
+Interactive:
+
+```bash
+codex
+```
+
+Non-interactive (one-shot):
+
+```bash
+codex exec "Refactor this function and add tests"
+```
+
+Every LLM call now flows through Eden AI — with built-in fallbacks, unified billing, and per-request cost visibility.
+
+---
+
+## Using specific models
+
+Switch model by editing `model` in `config.toml`, keyed by `provider/model`:
+
+| Model | Best for |
+|---|---|
+| `openai/gpt-4o-mini` | Fast and cheap (default above) |
+| `anthropic/claude-sonnet-4-5` | Balanced reasoning |
+| `anthropic/claude-opus-4-7` | Maximum capability |
+| `openai/gpt-5` | OpenAI's frontier model |
+| `google/gemini-2.5-pro` | Long context, multimodal |
+
+> Eden AI's smart router `@edenai` also works over the Responses API. If you use it, expect Codex to route to a reasoning-grade model automatically.
+
+## Troubleshooting
+
+### `wire_api = "chat" is no longer supported`
+Set `wire_api = "responses"` in your provider config (see step 3).
+
+### `401 Unauthorized`
+Verify your Eden AI key — grab a fresh one at [app.edenai.run → API Keys](https://app.edenai.run/). Confirm `EDENAI_API_KEY` is exported in the same shell.
+
+### `model not found`
+Use the `provider/model` format (e.g. `openai/gpt-4o-mini`). No bare model names.
+
+## Uninstall
+
+```bash
+npm uninstall -g @openai/codex
+rm -rf ~/.codex
+```
+
+---
+
+## Next steps
+
+- [Smart routing](https://www.edenai.co/docs/v3/llms/smart-routing)
+- [Fallbacks](https://www.edenai.co/docs/v3/llms/fallback)
+- [Eden AI pricing](https://www.edenai.co/pricing)
+- [Codex CLI GitHub](https://github.com/openai/codex)
diff --git a/continue.md b/continue.md
new file mode 100644
index 0000000..35380ef
--- /dev/null
+++ b/continue.md
@@ -0,0 +1,86 @@
+# Continue + Eden AI
+
+Connect [Continue](https://continue.dev) to Eden AI and route the `cn` agent (and the IDE extension) through **700+ models** with smart routing, automatic fallbacks, and a single API key.
+
+> Get your Eden AI key at [app.edenai.run](https://app.edenai.run/) → **API Keys**.
+
+---
+
+## 1. Install Continue
+
+- **IDE:** install the **Continue** extension from the VS Code or JetBrains marketplace.
+- **CLI:** install the `cn` terminal agent:
+
+```bash
+npm install -g @continuedev/cli
+cn --version
+```
+
+## 2. Add Eden AI in `config.yaml`
+
+Continue reads `~/.continue/config.yaml`. Add Eden AI as an OpenAI-compatible provider:
+
+```bash
+nano ~/.continue/config.yaml
+```
+
+```yaml
+models:
+  - name: GPT-4o mini (Eden AI)
+    provider: openai
+    model: openai/gpt-4o-mini
+    apiBase: https://api.edenai.run/v3
+    apiKey: YOUR_EDEN_AI_API_KEY
+    roles:
+      - chat
+      - edit
+  - name: Claude Haiku 4.5 (Eden AI)
+    provider: openai
+    model: anthropic/claude-haiku-4-5
+    apiBase: https://api.edenai.run/v3
+    apiKey: YOUR_EDEN_AI_API_KEY
+    roles:
+      - chat
+      - edit
+```
+
+> `provider: openai` selects the OpenAI-compatible driver; `apiBase` points it at Eden AI. Model ids use the `provider/model` format.
+
+## 3. Run
+
+Reload your IDE (or start `cn`) and pick an **Eden AI** model from the model selector. Every LLM call now flows through Eden AI — with built-in fallbacks, unified billing, and per-request cost visibility.
+
+---
+
+## Using specific models
+
+Add more entries under `models`, keyed by `provider/model`:
+
+| `model` | Best for |
+|---|---|
+| `openai/gpt-4o-mini` | Fast and cheap |
+| `anthropic/claude-sonnet-4-5` | Balanced reasoning |
+| `anthropic/claude-opus-4-7` | Maximum capability |
+| `openai/gpt-5` | OpenAI's frontier model |
+| `google/gemini-2.5-pro` | Long context, multimodal |
+
+## Troubleshooting
+
+### `401 Unauthorized`
+Verify your Eden AI key at [app.edenai.run → API Keys](https://app.edenai.run/).
+
+### `model not found`
+Use the `provider/model` format for `model`, and set `apiBase` to exactly `https://api.edenai.run/v3`.
+
+## Uninstall
+
+Remove the Continue extension from your IDE, or `npm uninstall -g @continuedev/cli`, then delete `~/.continue`.
+
+---
+
+## Next steps
+
+- [Smart routing](https://www.edenai.co/docs/v3/llms/smart-routing)
+- [Fallbacks](https://www.edenai.co/docs/v3/llms/fallback)
+- [Eden AI pricing](https://www.edenai.co/pricing)
+- [Continue docs](https://docs.continue.dev/)
diff --git a/crush.md b/crush.md
new file mode 100644
index 0000000..1d7d650
--- /dev/null
+++ b/crush.md
@@ -0,0 +1,123 @@
+# Crush + Eden AI
+
+Connect [Crush](https://github.com/charmbracelet/crush) to Eden AI and route Charm's glamourous coding TUI through **700+ models** with smart routing, automatic fallbacks, and a single API key.
+
+> Get your Eden AI key at [app.edenai.run](https://app.edenai.run/) → **API Keys**.
+
+---
+
+## 1. System dependencies
+
+On a fresh Linux machine (Ubuntu / Debian / WSL):
+
+```bash
+sudo apt update
+sudo apt install -y curl git
+```
+
+## 2. Install Crush
+
+```bash
+# Option A — via npm
+npm install -g @charmland/crush
+
+# Option B — Homebrew
+brew install charmbracelet/tap/crush
+
+# Verify
+crush --version
+```
+
+## 3. Add Eden AI as a custom provider
+
+Crush reads `crush.json` from your project directory (or `~/.config/crush/crush.json`). Add Eden AI as an OpenAI-compatible provider:
+
+```bash
+export EDENAI_API_KEY="YOUR_EDEN_AI_API_KEY"
+nano crush.json
+```
+
+```json
+{
+  "$schema": "https://charm.land/crush.json",
+  "providers": {
+    "edenai": {
+      "type": "openai",
+      "base_url": "https://api.edenai.run/v3",
+      "api_key": "$EDENAI_API_KEY",
+      "models": [
+        {
+          "id": "openai/gpt-4o-mini",
+          "name": "GPT-4o mini (Eden AI)",
+          "context_window": 128000,
+          "default_max_tokens": 4096
+        },
+        {
+          "id": "anthropic/claude-haiku-4-5",
+          "name": "Claude Haiku 4.5 (Eden AI)",
+          "context_window": 200000,
+          "default_max_tokens": 8192
+        }
+      ]
+    }
+  }
+}
+```
+
+> Use `type: "openai"` with a custom `base_url`. Every model you want to use must be listed under `models` with its `id` in `provider/model` format.
+
+## 4. Run Crush
+
+Interactive:
+
+```bash
+crush
+```
+
+Non-interactive (one-shot):
+
+```bash
+crush run -m edenai/openai/gpt-4o-mini "Explain what this repo does"
+```
+
+Every LLM call now flows through Eden AI — with built-in fallbacks, unified billing, and per-request cost visibility.
+
+---
+
+## Using specific models
+
+Add any model to the `models` array, then select it with `-m edenai/<provider/model>`:
+
+| Model `id` | Best for |
+|---|---|
+| `openai/gpt-4o-mini` | Fast and cheap |
+| `anthropic/claude-sonnet-4-5` | Balanced reasoning |
+| `anthropic/claude-opus-4-7` | Maximum capability |
+| `openai/gpt-5` | OpenAI's frontier model |
+| `google/gemini-2.5-pro` | Long context, multimodal |
+
+> Crush's default catalog comes from Charm's **Catwalk** registry. Use a concrete `provider/model`; the `@edenai` smart-router alias is not currently supported through Crush's custom-provider path.
+
+## Troubleshooting
+
+### `401 Unauthorized`
+Verify your Eden AI key at [app.edenai.run → API Keys](https://app.edenai.run/). Confirm `EDENAI_API_KEY` is exported.
+
+### `model not found`
+The model must be listed in the `models` array with its `id` in `provider/model` format, and selected as `edenai/<provider/model>`.
+
+## Uninstall
+
+```bash
+npm uninstall -g @charmland/crush
+rm -rf ~/.config/crush
+```
+
+---
+
+## Next steps
+
+- [Smart routing](https://www.edenai.co/docs/v3/llms/smart-routing)
+- [Fallbacks](https://www.edenai.co/docs/v3/llms/fallback)
+- [Eden AI pricing](https://www.edenai.co/pricing)
+- [Crush GitHub](https://github.com/charmbracelet/crush)
diff --git a/goose.md b/goose.md
new file mode 100644
index 0000000..3cf25d0
--- /dev/null
+++ b/goose.md
@@ -0,0 +1,114 @@
+# Goose + Eden AI
+
+Connect [Goose](https://block.github.io/goose/) to Eden AI and route Block's open-source agent through **700+ models** with smart routing, automatic fallbacks, and a single API key.
+
+> Get your Eden AI key at [app.edenai.run](https://app.edenai.run/) → **API Keys**.
+
+---
+
+## 1. System dependencies
+
+On a fresh Linux machine (Ubuntu / Debian / WSL):
+
+```bash
+sudo apt update
+sudo apt install -y curl bzip2
+```
+
+## 2. Install Goose
+
+```bash
+curl -fsSL https://github.com/block/goose/releases/download/stable/download_cli.sh | bash
+
+# Verify
+goose --version
+```
+
+## 3. Configure Eden AI (env variables)
+
+Goose speaks to any OpenAI-compatible endpoint. The simplest setup uses environment variables:
+
+```bash
+export GOOSE_PROVIDER="openai"
+export GOOSE_MODEL="openai/gpt-4o-mini"
+export OPENAI_HOST="https://api.edenai.run"
+export OPENAI_BASE_PATH="v3/chat/completions"
+export OPENAI_API_KEY="YOUR_EDEN_AI_API_KEY"
+```
+
+> **Split host and path.** Set `OPENAI_HOST` to the root (`https://api.edenai.run`) and `OPENAI_BASE_PATH` to `v3/chat/completions`. This split is what Goose expects for a custom OpenAI-compatible endpoint.
+
+### Alternative — custom provider file
+
+Create `~/.config/goose/custom_providers/edenai.json`:
+
+```json
+{
+  "name": "edenai",
+  "display_name": "Eden AI",
+  "engine": "openai",
+  "api_key_env": "EDENAI_API_KEY",
+  "base_url": "https://api.edenai.run",
+  "base_path": "v3/chat/completions",
+  "models": ["openai/gpt-4o-mini", "anthropic/claude-haiku-4-5"]
+}
+```
+
+Then run `goose configure` and select **Eden AI**.
+
+## 4. Run
+
+Interactive:
+
+```bash
+goose session
+```
+
+Non-interactive (one-shot):
+
+```bash
+goose run --no-session -t "Summarize the README in this repo"
+```
+
+Every LLM call now flows through Eden AI — with built-in fallbacks, unified billing, and per-request cost visibility.
+
+---
+
+## Using specific models
+
+Set `GOOSE_MODEL` (or pick in the wizard), keyed by `provider/model`:
+
+| Model | Best for |
+|---|---|
+| `openai/gpt-4o-mini` | Fast and cheap |
+| `anthropic/claude-sonnet-4-5` | Balanced reasoning |
+| `anthropic/claude-opus-4-7` | Maximum capability |
+| `openai/gpt-5` | OpenAI's frontier model |
+| `google/gemini-2.5-pro` | Long context, multimodal |
+
+## Troubleshooting
+
+### `401 Unauthorized`
+Verify your Eden AI key at [app.edenai.run → API Keys](https://app.edenai.run/).
+
+### `404` / wrong path
+Make sure `OPENAI_HOST` is the root and `OPENAI_BASE_PATH` is `v3/chat/completions` — not the full URL in a single variable.
+
+### Keyring errors on headless machines
+Set `GOOSE_DISABLE_KEYRING=1` before running.
+
+## Uninstall
+
+```bash
+rm -f ~/.local/bin/goose
+rm -rf ~/.config/goose
+```
+
+---
+
+## Next steps
+
+- [Smart routing](https://www.edenai.co/docs/v3/llms/smart-routing)
+- [Fallbacks](https://www.edenai.co/docs/v3/llms/fallback)
+- [Eden AI pricing](https://www.edenai.co/pricing)
+- [Goose docs](https://block.github.io/goose/)
diff --git a/openhands.md b/openhands.md
new file mode 100644
index 0000000..4975db9
--- /dev/null
+++ b/openhands.md
@@ -0,0 +1,88 @@
+# OpenHands + Eden AI
+
+Connect [OpenHands](https://github.com/All-Hands-AI/OpenHands) to Eden AI and route the most autonomous open-source agent through **700+ models** with smart routing, automatic fallbacks, and a single API key.
+
+> Get your Eden AI key at [app.edenai.run](https://app.edenai.run/) → **API Keys**.
+
+---
+
+## 1. System dependencies
+
+OpenHands runs in a sandboxed runtime. You need **Docker** (for the runtime) and optionally Python 3.12+ for the CLI:
+
+```bash
+sudo apt update
+sudo apt install -y docker.io python3 python3-venv
+```
+
+## 2. Install OpenHands
+
+```bash
+# CLI (recommended for terminal use)
+uvx --python 3.12 --from openhands-ai openhands
+
+# Or GUI in Docker — see the official quickstart:
+# https://docs.all-hands.dev/usage/getting-started
+```
+
+## 3. Configure Eden AI as the LLM
+
+OpenHands talks to models through LiteLLM, which supports OpenAI-compatible endpoints.
+
+**Via the UI:** open **Settings → LLM**, toggle **Advanced**, and set:
+
+- **Custom Model:** `openai/openai/gpt-4o-mini`
+- **Base URL:** `https://api.edenai.run/v3`
+- **API Key:** `YOUR_EDEN_AI_API_KEY`
+
+**Via `config.toml`:**
+
+```toml
+[llm]
+model = "openai/openai/gpt-4o-mini"
+base_url = "https://api.edenai.run/v3"
+api_key = "YOUR_EDEN_AI_API_KEY"
+```
+
+> **Why the double prefix?** LiteLLM's leading `openai/` selects the OpenAI-compatible driver and is stripped before the request; the remaining `openai/gpt-4o-mini` is the Eden AI model id. For Claude: `openai/anthropic/claude-haiku-4-5`.
+
+## 4. Run
+
+Start OpenHands and give it a task. Every LLM call now flows through Eden AI — with built-in fallbacks, unified billing, and per-request cost visibility.
+
+---
+
+## Using specific models
+
+The model string is `openai/<eden-provider>/<eden-model>`:
+
+| Model value | Best for |
+|---|---|
+| `openai/openai/gpt-4o-mini` | Fast and cheap |
+| `openai/anthropic/claude-sonnet-4-5` | Balanced reasoning |
+| `openai/anthropic/claude-opus-4-7` | Maximum capability |
+| `openai/openai/gpt-5` | OpenAI's frontier model |
+| `openai/google/gemini-2.5-pro` | Long context, multimodal |
+
+> Use a concrete `provider/model`; the `@edenai` smart-router alias is not recommended through LiteLLM.
+
+## Troubleshooting
+
+### `401 Unauthorized`
+Verify your Eden AI key at [app.edenai.run → API Keys](https://app.edenai.run/).
+
+### `model not found` / `LLM Provider NOT provided`
+Keep the double prefix (`openai/` + Eden AI `provider/model`) and set the Base URL to exactly `https://api.edenai.run/v3`.
+
+## Uninstall
+
+Stop and remove the Docker containers/images, or `uv tool uninstall openhands-ai` if you installed the CLI.
+
+---
+
+## Next steps
+
+- [Smart routing](https://www.edenai.co/docs/v3/llms/smart-routing)
+- [Fallbacks](https://www.edenai.co/docs/v3/llms/fallback)
+- [Eden AI pricing](https://www.edenai.co/pricing)
+- [OpenHands docs](https://docs.all-hands.dev/)
diff --git a/pi.md b/pi.md
new file mode 100644
index 0000000..23ae256
--- /dev/null
+++ b/pi.md
@@ -0,0 +1,122 @@
+# Pi + Eden AI
+
+Connect [Pi](https://github.com/earendil-works/pi) to Eden AI and route the lean, hackable coding agent through **700+ models** with smart routing, automatic fallbacks, and a single API key.
+
+> Get your Eden AI key at [app.edenai.run](https://app.edenai.run/) → **API Keys**.
+
+---
+
+## 1. Install Node.js
+
+Pi requires **Node.js 22+**.
+
+```bash
+node --version   # must be v22.x or newer
+```
+
+If it's older, install via [nvm](https://github.com/nvm-sh/nvm):
+
+```bash
+curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.1/install.sh | bash
+nvm install 22 && nvm use 22
+```
+
+## 2. Install Pi
+
+```bash
+npm install -g @earendil-works/pi-coding-agent
+
+# Verify
+pi --version
+```
+
+## 3. Add Eden AI as a custom provider
+
+Pi reads custom providers from `~/.pi/agent/models.json`. Add Eden AI as an OpenAI-compatible provider:
+
+```bash
+export EDENAI_API_KEY="YOUR_EDEN_AI_API_KEY"
+mkdir -p ~/.pi/agent
+nano ~/.pi/agent/models.json
+```
+
+```json
+{
+  "providers": {
+    "edenai": {
+      "baseUrl": "https://api.edenai.run/v3",
+      "api": "openai-completions",
+      "apiKey": "$EDENAI_API_KEY",
+      "models": [
+        { "id": "openai/gpt-4o-mini", "name": "GPT-4o mini (Eden AI)" },
+        { "id": "anthropic/claude-haiku-4-5", "name": "Claude Haiku 4.5 (Eden AI)" }
+      ]
+    }
+  }
+}
+```
+
+The `apiKey` field supports `$ENV_VAR` interpolation, so the key stays out of the file. Confirm the models loaded:
+
+```bash
+pi --list-models edenai
+```
+
+## 4. Run Pi
+
+Interactive:
+
+```bash
+pi --provider edenai --model openai/gpt-4o-mini
+```
+
+Non-interactive (one-shot):
+
+```bash
+pi -p --model edenai/openai/gpt-4o-mini "Explain what this repo does"
+```
+
+Every LLM call now flows through Eden AI — with built-in fallbacks, unified billing, and per-request cost visibility.
+
+---
+
+## Using specific models
+
+Add any model to the `models` array, then select it as `edenai/<provider/model>`:
+
+| Model `id` | Best for |
+|---|---|
+| `openai/gpt-4o-mini` | Fast and cheap |
+| `anthropic/claude-sonnet-4-5` | Balanced reasoning |
+| `anthropic/claude-opus-4-7` | Maximum capability |
+| `openai/gpt-5` | OpenAI's frontier model |
+| `google/gemini-2.5-pro` | Long context, multimodal |
+
+> Some OpenAI-compatible reasoning models don't accept the `developer` role. If you hit an error with a reasoning model, add `"compat": { "supportsDeveloperRole": false, "supportsReasoningEffort": false }` at the provider or model level in `models.json`.
+
+## Troubleshooting
+
+### `No models available`
+Check `~/.pi/agent/models.json` is valid JSON and that `EDENAI_API_KEY` is exported, then run `pi --list-models edenai`.
+
+### `401 Unauthorized`
+Verify your Eden AI key at [app.edenai.run → API Keys](https://app.edenai.run/).
+
+### `model not found`
+Use the `edenai/<provider>/<model>` format (e.g. `edenai/anthropic/claude-haiku-4-5`), and make sure the model is listed under `models`.
+
+## Uninstall
+
+```bash
+npm uninstall -g @earendil-works/pi-coding-agent
+rm -rf ~/.pi
+```
+
+---
+
+## Next steps
+
+- [Smart routing](https://www.edenai.co/docs/v3/llms/smart-routing)
+- [Fallbacks](https://www.edenai.co/docs/v3/llms/fallback)
+- [Eden AI pricing](https://www.edenai.co/pricing)
+- [Pi GitHub](https://github.com/earendil-works/pi)
diff --git a/qwen-code.md b/qwen-code.md
new file mode 100644
index 0000000..0cef378
--- /dev/null
+++ b/qwen-code.md
@@ -0,0 +1,100 @@
+# Qwen Code + Eden AI
+
+Connect [Qwen Code](https://github.com/QwenLM/qwen-code) to Eden AI and route Alibaba's terminal agent through **700+ models** with smart routing, automatic fallbacks, and a single API key.
+
+> Get your Eden AI key at [app.edenai.run](https://app.edenai.run/) → **API Keys**.
+
+---
+
+## 1. Install Node.js
+
+Qwen Code requires **Node.js 22+**.
+
+```bash
+# Check your version
+node --version   # must be v22.x or newer
+```
+
+If it's older, install via [nvm](https://github.com/nvm-sh/nvm):
+
+```bash
+curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.1/install.sh | bash
+nvm install 22 && nvm use 22
+```
+
+## 2. Install Qwen Code
+
+```bash
+npm install -g @qwen-code/qwen-code
+
+# Verify
+qwen --version
+```
+
+## 3. Configure Eden AI (env variables)
+
+Qwen Code (a fork of Gemini CLI) reads OpenAI-compatible environment variables:
+
+```bash
+export OPENAI_API_KEY="YOUR_EDEN_AI_API_KEY"
+export OPENAI_BASE_URL="https://api.edenai.run/v3"
+export OPENAI_MODEL="openai/gpt-4o-mini"
+```
+
+To make it persistent, add these to a project `.env` file or your shell profile.
+
+## 4. Run
+
+Interactive:
+
+```bash
+qwen
+```
+
+Non-interactive (one-shot):
+
+```bash
+qwen -p "Explain what this project does"
+```
+
+Every LLM call now flows through Eden AI — with built-in fallbacks, unified billing, and per-request cost visibility.
+
+---
+
+## Using specific models
+
+Change `OPENAI_MODEL`, keyed by `provider/model`:
+
+| Model | Best for |
+|---|---|
+| `openai/gpt-4o-mini` | Fast and cheap |
+| `anthropic/claude-sonnet-4-5` | Balanced reasoning |
+| `anthropic/claude-opus-4-7` | Maximum capability |
+| `openai/gpt-5` | OpenAI's frontier model |
+| `google/gemini-2.5-pro` | Long context, multimodal |
+
+## Troubleshooting
+
+### `ReferenceError: File is not defined`
+Your Node.js is too old. Qwen Code needs **Node 22+** — upgrade via nvm (step 1).
+
+### `401 Unauthorized`
+Verify your Eden AI key at [app.edenai.run → API Keys](https://app.edenai.run/).
+
+### `model not found`
+Use the `provider/model` format for `OPENAI_MODEL`, and set `OPENAI_BASE_URL` to exactly `https://api.edenai.run/v3`.
+
+## Uninstall
+
+```bash
+npm uninstall -g @qwen-code/qwen-code
+```
+
+---
+
+## Next steps
+
+- [Smart routing](https://www.edenai.co/docs/v3/llms/smart-routing)
+- [Fallbacks](https://www.edenai.co/docs/v3/llms/fallback)
+- [Eden AI pricing](https://www.edenai.co/pricing)
+- [Qwen Code GitHub](https://github.com/QwenLM/qwen-code)