rodiumai

Official Dart/Flutter SDK for the Rodium AI API — access GPT, Claude, Gemini, DeepSeek and more with RODI credits and Mobile Money top-ups.

OpenAI-compatible REST API — same endpoints and payloads as documented at rodiumai.io/docs.

Links

Resource	URL
pub.dev	pub.dev/packages/rodiumai
Source code	github.com/lecodeur228/rodiumai
API documentation	rodiumai.io/docs
Dashboard & API keys	rodiumai.io/dashboard
Model catalogue	rodiumai.io/models

Installation (Flutter)

Add the package to your Flutter or Dart project:

flutter pub add rodiumai
# ou pour un projet Dart pur :
dart pub add rodiumai

Requirements: Dart SDK >=3.0.0. No Flutter dependency — the package works in CLI, server, and Flutter apps alike.

Configuration

1. Get an API key

Create a free account on rodiumai.io/dashboard
Top up RODI credits (Mobile Money: MTN, Orange, Wave, Moov, Airtel…)
Generate an API key (prefix rd_sk_… or rdk_…)

2. Provide the key to your app

Option A — environment variable (recommended for dev/CI):

export RODIUMAI_API_KEY="rd_sk_votre_cle"
dart run example/main.dart

Option B — --dart-define (Flutter run/build):

flutter run --dart-define=RODIUMAI_API_KEY=rd_sk_votre_cle

const apiKey = String.fromEnvironment('RODIUMAI_API_KEY');
if (apiKey.isEmpty) throw StateError('RODIUMAI_API_KEY manquante');

Option C — flutter_dotenv (.env file, never commit secrets):

# pubspec.yaml
dependencies:
  rodiumai: ^0.1.0
  flutter_dotenv: ^5.2.0

flutter:
  assets:
    - .env

import 'package:flutter_dotenv/flutter_dotenv.dart';

await dotenv.load();
final client = RodiumAIClient(apiKey: dotenv.env['RODIUMAI_API_KEY']!);

3. Create the client

import 'package:rodiumai/rodiumai.dart';

final client = RodiumAIClient(
  apiKey: 'rd_sk_votre_cle',
  // logRequests: true,  // debug HTTP (dev only)
);

Multilingual (optional)

Language is never required. Without configuration, the SDK uses English error hints and the AI follows the user's message language.

Optional — localized SDK error hints

final client = RodiumAIClient(
  apiKey: apiKey,
  locale: 'fr', // optional — only affects exception hint text
);

Optional — force AI responses in app language

// Only when you want the model to always reply in a specific language
final res = await client
    .language('fr') // optional fluent method
    .model('openai/gpt-4o')
    .chat('Bonjour !');

Flutter example (optional):

final locale = Localizations.localeOf(context).languageCode;
await client.language(locale).model(modelId).chat('Hello');

Supported locales: en (default), fr, es.

Unlike a static enum, models are fetched live from the API so your app always reflects the current catalogue and RODI pricing metadata.

final catalogue = await client.models();

print('${catalogue.length} modèles');           // ex: 42
print(catalogue.providerPrefixes);              // [anthropic, deepseek, google, …]
print(catalogue.chatModels.map((m) => m.id));   // modèles compatibles chat

// Filtres
final openai = catalogue.byProvider('openai');
final anthropic = catalogue.byProvider('anthropic');
final longCtx = catalogue.withContextAtLeast(500000);
final gpt4o = catalogue.findById('openai/gpt-4o');

// Propriétés d'un modèle
for (final m in catalogue.chatModels) {
  print('${m.displayName} — ${m.contextWindow} tokens — chat: ${m.supportsChatCompletion}');
}

Reference: GET /v1/models

Quick start

Chat with a string

final res = await client.chat('Bonjour !');
print(res.content);
print('Tokens: ${res.totalTokens}');

Chat with fluent builder + dynamic model

final catalogue = await client.models();
final modelId = catalogue.findById('openai/gpt-4o')?.id
    ?? catalogue.chatModels.first.id;

final res = await client
    .model(modelId)
    .temperature(0.7)
    .maxTokens(256)
    .systemPrompt('Réponds toujours en français.')
    .chat('Présente RodiumAI en 2 phrases.');

print(res.content);

Multi-turn conversation

final messages = [
  const ChatMessage.user('Quel est le capital du Togo ?'),
  const ChatMessage.assistant('Lomé est la capitale du Togo.'),
  const ChatMessage.user('Et sa population ?'),
];

final res = await client
    .model('anthropic/claude-haiku-4-5-20251001')
    .chat(messages);

Streaming SSE

Per Streaming guide: lines data: {json}, end with data: [DONE].

await for (final delta in client.model('openai/gpt-4o-mini').stream('Raconte une histoire courte.')) {
  stdout.write(delta.content);
}

Flutter integration

StatefulWidget — chat screen

import 'package:flutter/material.dart';
import 'package:rodiumai/rodiumai.dart';

class ChatScreen extends StatefulWidget {
  const ChatScreen({super.key, required this.client});
  final RodiumAIClient client;

  @override
  State<ChatScreen> createState() => _ChatScreenState();
}

class _ChatScreenState extends State<ChatScreen> {
  final _controller = TextEditingController();
  String _response = '';
  bool _loading = false;
  List<ModelInfo> _models = [];
  String? _selectedModelId;

  @override
  void initState() {
    super.initState();
    _loadModels();
  }

  Future<void> _loadModels() async {
    try {
      final catalogue = await widget.client.models();
      setState(() {
        _models = catalogue.chatModels;
        _selectedModelId = _models.isNotEmpty ? _models.first.id : null;
      });
    } on RodiumAIException catch (e) {
      if (mounted) {
        ScaffoldMessenger.of(context).showSnackBar(
          SnackBar(content: Text('Erreur modèles: ${e.message}')),
        );
      }
    }
  }

  Future<void> _send() async {
    final prompt = _controller.text.trim();
    if (prompt.isEmpty || _selectedModelId == null) return;

    setState(() {
      _loading = true;
      _response = '';
    });

    try {
      final res = await widget.client
          .model(_selectedModelId!)
          .temperature(0.7)
          .maxTokens(512)
          .chat(prompt);
      setState(() => _response = res.content);
    } on InsufficientCreditsException {
      if (mounted) {
        ScaffoldMessenger.of(context).showSnackBar(
          const SnackBar(content: Text('Crédits RODI insuffisants — rechargez sur rodiumai.io')),
        );
      }
    } on RodiumAIException catch (e) {
      setState(() => _response = 'Erreur (${e.statusCode}): ${e.message}');
    } finally {
      setState(() => _loading = false);
    }
  }

  @override
  Widget build(BuildContext context) {
    return Scaffold(
      appBar: AppBar(title: const Text('RodiumAI Chat')),
      body: Padding(
        padding: const EdgeInsets.all(16),
        child: Column(
          children: [
            if (_models.isNotEmpty)
              DropdownButtonFormField<String>(
                value: _selectedModelId,
                decoration: const InputDecoration(labelText: 'Modèle'),
                items: _models
                    .map((m) => DropdownMenuItem(
                          value: m.id,
                          child: Text('${m.displayName} (${m.providerPrefix})'),
                        ))
                    .toList(),
                onChanged: (v) => setState(() => _selectedModelId = v),
              ),
            const SizedBox(height: 12),
            TextField(
              controller: _controller,
              decoration: const InputDecoration(
                labelText: 'Message',
                border: OutlineInputBorder(),
              ),
              maxLines: 3,
            ),
            const SizedBox(height: 12),
            FilledButton(
              onPressed: _loading ? null : _send,
              child: _loading
                  ? const SizedBox(
                      width: 20,
                      height: 20,
                      child: CircularProgressIndicator(strokeWidth: 2),
                    )
                  : const Text('Envoyer'),
            ),
            const SizedBox(height: 16),
            Expanded(
              child: SingleChildScrollView(
                child: Text(_response.isEmpty ? '…' : _response),
              ),
            ),
          ],
        ),
      ),
    );
  }

  @override
  void dispose() {
    _controller.dispose();
    super.dispose();
  }
}

Streaming in Flutter (typing effect)

class StreamingChat extends StatefulWidget {
  const StreamingChat({super.key, required this.client, required this.modelId});
  final RodiumAIClient client;
  final String modelId;

  @override
  State<StreamingChat> createState() => _StreamingChatState();
}

class _StreamingChatState extends State<StreamingChat> {
  final _buffer = StringBuffer();

  Future<void> _stream(String prompt) async {
    _buffer.clear();
    setState(() {});

    await for (final delta in widget.client.model(widget.modelId).stream(prompt)) {
      setState(() => _buffer.write(delta.content));
    }
  }

  @override
  Widget build(BuildContext context) {
    return Column(
      children: [
        FilledButton(
          onPressed: () => _stream('Compte de 1 à 5.'),
          child: const Text('Stream'),
        ),
        Expanded(child: SingleChildScrollView(child: Text(_buffer.toString()))),
      ],
    );
  }
}

Provider filter tabs

Widget buildProviderTabs(ModelCollection catalogue, void Function(String) onSelect) {
  final providers = catalogue.providerPrefixes;
  return Wrap(
    spacing: 8,
    children: providers.map((prefix) {
      final count = catalogue.byProvider(prefix).chatModels.length;
      return ActionChip(
        label: Text('$prefix ($count)'),
        onPressed: () {
          final chat = catalogue.byProvider(prefix).chatModels;
          if (chat.isNotEmpty) onSelect(chat.first.id);
        },
      );
    }).toList(),
  
  );
}

Error handling

Mapped per API errors:

HTTP	Exception	Action in Flutter
401	`UnauthorizedException`	Check API key / re-login dashboard
402	`InsufficientCreditsException`	Show top-up link to dashboard
429	`RateLimitException`	Retry with exponential backoff
422	`ValidationException`	Fix `model` or `messages` payload
500+	`RodiumAIException`	Retry once, then support

try {
  await client.chat('Hello');
} on UnauthorizedException catch (e) {
  // Clé invalide
} on InsufficientCreditsException catch (e) {
  // Solde RODI insuffisant
} on RateLimitException catch (e) {
  // Trop de requêtes
} on ValidationException catch (e) {
  // Payload invalide
} on RodiumAIException catch (e) {
  print('${e.statusCode}: ${e.message}');
  print(e.responseBody); // JSON brut si présent
}

API reference

`RodiumAIClient`

Method	Returns	Description
`chat(messages)`	`Future<ChatResponse>`	Non-streaming completion
`stream(messages)`	`Stream<ChatStreamDelta>`	SSE streaming
`models()`	`Future<ModelCollection>`	Live catalogue from API
`model(id)`	`RodiumAIClient`	Set model ID (fluent)
`temperature(n)`	`RodiumAIClient`	0.0–2.0
`topP(n)`	`RodiumAIClient`	0.0–1.0
`maxTokens(n)`	`RodiumAIClient`	Max output tokens
`systemPrompt(s)`	`RodiumAIClient`	Prepended system message
`stop(list)`	`RodiumAIClient`	Stop sequences

messages accepts: String, List<ChatMessage>, or List<Map<String, dynamic>>.

All fluent methods return a new immutable instance — safe to reuse the base client.

`ModelCollection`

Member	Description
`models`	Full list from API
`chatModels`	Text/chat-capable models only
`ids`	All model ID strings
`findById(id)`	Lookup by exact ID
`byProvider(prefix)`	Filter e.g. `"openai"` (dynamic from API)
`withContextAtLeast(n)`	Long-context models
`reasoningModels`	o1, R1, M1…
`providerPrefixes`	Distinct providers in response

`ModelInfo`

Property	Description
`id`	API model slug (`openai/gpt-4o`)
`contextWindow`	Context size in tokens
`outputModalities`	`text`, `image`, `embedding`…
`displayName`	Human-readable label
`supportsChatCompletion`	Usable with `chat()` / `stream()`
`isLongContext`	≥ 200K tokens
`isReasoningModel`	Chain-of-thought models
`provider` / `providerPrefix`	Provider metadata

`ChatMessage` constructors

ChatMessage.user('Hello');
ChatMessage.assistant('Previous reply');
ChatMessage.system('You are a helpful assistant.');

Endpoints implemented

API	SDK	Docs
`POST /v1/chat/completions`	`chat()`, `stream()`	Chat completions
`GET /v1/models`	`models()`	List models

Base URL: https://api.rodiumai.io/v1 (fixed — not configurable)
Auth: Authorization: Bearer {RODIUMAI_API_KEY}

Example app

cd rodiumai
export RODIUMAI_API_KEY="rd_sk_votre_cle"
dart run example/main.dart

Testing & development

See CONTRIBUTING.md for project structure and PR guidelines.

dart pub get
dart run build_runner build
dart analyze --fatal-infos
dart test
dart pub publish --dry-run

License

MIT — see LICENSE.

rodiumai

Links

Installation (Flutter)

Configuration

1. Get an API key

2. Provide the key to your app

3. Create the client

Multilingual (optional)

Optional — localized SDK error hints

Optional — force AI responses in app language

Quick start

Chat with a string

Chat with fluent builder + dynamic model

Multi-turn conversation

Streaming SSE

Flutter integration

StatefulWidget — chat screen

Streaming in Flutter (typing effect)

Provider filter tabs

Error handling

API reference

`RodiumAIClient`

`ModelCollection`

`ModelInfo`

`ChatMessage` constructors

Endpoints implemented

Example app

Testing & development

License

Libraries

rodiumai package

rodiumai

Links

Installation (Flutter)

Configuration

1. Get an API key

2. Provide the key to your app

3. Create the client

Multilingual (optional)

Optional — localized SDK error hints

Optional — force AI responses in app language

Quick start

Chat with a string

Chat with fluent builder + dynamic model

Multi-turn conversation

Streaming SSE

Flutter integration

StatefulWidget — chat screen

Streaming in Flutter (typing effect)

Provider filter tabs

Error handling

API reference

RodiumAIClient

ModelCollection

ModelInfo

ChatMessage constructors

Endpoints implemented

Example app

Testing & development

License

Libraries

rodiumai package

`RodiumAIClient`

`ModelCollection`

`ModelInfo`

`ChatMessage` constructors