Chat Completions — API Reference

curl -sS -X POST "https://api.dev.dodil.io/v1/chat/completions" \
  -H "Authorization: Bearer $DODIL_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kimi-k2.5",
    "messages": [
      { "role": "system", "content": "You are a concise assistant." },
      { "role": "user", "content": "Summarize the CAP theorem in one sentence." }
    ],
    "max_tokens": 256,
    "temperature": 0.7,
    "top_p": 0.95,
    "response_format": { "type": "text" }
  }'

rpc ChatCompletion(ChatCompletionRequest) returns (ChatCompletionResponse);
 
message ChatCompletionRequest {
  string model = 1;
  repeated ChatMessage messages = 2;
  int32 max_tokens = 3;
  double temperature = 4;
  double top_p = 5;
  double frequency_penalty = 6;
  double presence_penalty = 7;
  repeated string stop = 8;
  int32 n = 9;
  bool stream = 10;
  int64 seed = 11;
  repeated google.protobuf.Value tools = 12;     // OpenAI tool definitions
  google.protobuf.Value tool_choice = 13;         // "none" | "auto" | "required" | { ... }
  google.protobuf.Value response_format = 14;     // { "type": "text" | "json_object" | "json_schema" }
  bool logprobs = 15;
  int32 top_logprobs = 16;
  bool parallel_tool_calls = 17;
  int32 max_completion_tokens = 18;
}
 
message ChatMessage {
  string role = 1;                                // system | user | assistant | tool
  google.protobuf.Value content = 2;              // string OR multimodal array
  string name = 3;
  string reasoning_content = 4;
  repeated google.protobuf.Value tool_calls = 5;
  string tool_call_id = 6;
}

grpcurl \
  -H "Authorization: Bearer $DODIL_TOKEN" \
  -d '{
    "model": "kimi-k2.5",
    "messages": [
      { "role": "system", "content": "You are a concise assistant." },
      { "role": "user", "content": "Summarize the CAP theorem in one sentence." }
    ],
    "max_tokens": 256,
    "temperature": 0.7,
    "top_p": 0.95,
    "response_format": { "type": "text" }
  }' \
  $IGNITE_GRPC \
  dodil.ignite.v1.ModelService/ChatCompletion

{
  "id": "chatcmpl-0a1b2c3d",
  "object": "chat.completion",
  "created": 1748390400,
  "model": "kimi-k2.5",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "A distributed store can guarantee at most two of consistency, availability, and partition tolerance at once."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": { "prompt_tokens": 28, "completion_tokens": 19, "total_tokens": 47 }
}

message ChatCompletionResponse {
  string id = 1;
  string object = 2;                              // "chat.completion"
  int64 created = 3;
  string model = 4;
  repeated Choice choices = 5;
  Usage usage = 6;
}
 
message Choice {
  int32 index = 1;
  ChatMessage message = 2;
  string finish_reason = 3;                       // stop | length | tool_calls | ...
}
 
message Usage {
  int32 prompt_tokens = 1;
  int32 completion_tokens = 2;
  int32 total_tokens = 3;
}

curl -sS -N -X POST "https://api.dev.dodil.io/v1/chat/completions/stream" \
  -H "Authorization: Bearer $DODIL_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kimi-k2.5",
    "messages": [
      { "role": "user", "content": "Write a haiku about distributed systems." }
    ],
    "stream": true,
    "max_tokens": 128
  }'

rpc StreamChatCompletion(ChatCompletionRequest) returns (stream ChatCompletionChunk);
 
message ChatCompletionChunk {
  string id = 1;
  string object = 2;                              // "chat.completion.chunk"
  int64 created = 3;
  string model = 4;
  repeated ChunkChoice choices = 5;
}
 
message ChunkChoice {
  int32 index = 1;
  ChatMessage delta = 2;                          // incremental fragment
  string finish_reason = 3;                       // set on the final chunk
}

grpcurl \
  -H "Authorization: Bearer $DODIL_TOKEN" \
  -d '{
    "model": "kimi-k2.5",
    "messages": [
      { "role": "user", "content": "Write a haiku about distributed systems." }
    ],
    "stream": true,
    "max_tokens": 128
  }' \
  $IGNITE_GRPC \
  dodil.ignite.v1.ModelService/StreamChatCompletion