Skip to content

Add the endpoints /api/tags and /api/chat #13659

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
May 21, 2025
Merged
Changes from 2 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
71 changes: 68 additions & 3 deletions tools/server/server.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -3702,6 +3702,7 @@ int main(int argc, char ** argv) {
"/health",
"/models",
"/v1/models",
"/api/tags"
};

// If API key is not set, skip validation
Expand Down Expand Up @@ -3740,7 +3741,7 @@ int main(int argc, char ** argv) {
if (req.path == "/" || tmp.back() == "html") {
res.set_content(reinterpret_cast<const char*>(loading_html), loading_html_len, "text/html; charset=utf-8");
res.status = 503;
} else if (req.path == "/models" || req.path == "/v1/models") {
} else if (req.path == "/models" || req.path == "/v1/models" || req.path == "/api/tags") {
// allow the models endpoint to be accessed during loading
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

during loading, common_chat_templates_source call will fail

Copy link
Contributor Author

@R-Dson R-Dson May 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you want me to remove that endpoint from the if-case?

Edit:
The current code does not use common_chat_templates_source in that endpoint after the latest changes.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

hmm I think it's ok to keep this then

return true;
} else {
Expand Down Expand Up @@ -4076,7 +4077,21 @@ int main(int argc, char ** argv) {
res_ok(res, {{ "success", true }});
};

const auto handle_api_show = [&ctx_server, &res_ok](const httplib::Request &, httplib::Response & res) {
const auto handle_api_show = [&ctx_server, &state, &res_ok](const httplib::Request &, httplib::Response & res) {
server_state current_state = state.load();
const auto* model = llama_get_model(ctx_server.ctx);

// Get basic model info
char arch_buf[64] = {0};
char param_size_buf[64] = {0};
llama_model_meta_val_str(model, "general.architecture", arch_buf, sizeof(arch_buf));
llama_model_meta_val_str(model, "general.parameter_count", param_size_buf, sizeof(param_size_buf));

json model_meta = nullptr;
if (current_state == SERVER_STATE_READY) {
model_meta = ctx_server.model_meta();
}

json data = {
{
"template", common_chat_templates_source(ctx_server.chat_templates.get()),
Expand All @@ -4086,6 +4101,19 @@ int main(int argc, char ** argv) {
{ "llama.context_length", ctx_server.slots.back().n_ctx, },
}
},
{"modelfile", ""}, // Specific to ollama and does not seem to be needed
{"parameters", ""}, // TODO: add parameters
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Does the API client expect this to be a string or an object?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

modelfile is a string. Parameters i am not sure, but it seems to be enough as is.

{"template", common_chat_templates_source(ctx_server.chat_templates.get())},
{"details", {
{"parent_model", ""}, // TODO: add parent model if available
{"format", "gguf"},
{"family", arch_buf},
{"families", {arch_buf}},
{"parameter_size", param_size_buf},
{"quantization_level", ""} // TODO: add quantization level if available
}},
{"model_info", model_meta},
{"capabilities", {"completion"}} // TODO: add other capabilities if available
};

res_ok(res, data);
Expand Down Expand Up @@ -4409,8 +4437,43 @@ int main(int argc, char ** argv) {
if (current_state == SERVER_STATE_READY) {
model_meta = ctx_server.model_meta();
}
// Get file metadata
struct stat file_stat;
stat(params.model.path.c_str(), &file_stat);

// Convert modified time to ISO 8601
char modified_buf[64];
strftime(modified_buf, sizeof(modified_buf), "%Y-%m-%dT%H:%M:%S%z", localtime(&file_stat.st_mtime));

const auto* model = llama_get_model(ctx_server.ctx);
char arch_buf[64] = {0};
char param_size_buf[64] = {0};
llama_model_meta_val_str(model, "general.architecture", arch_buf, sizeof(arch_buf));
llama_model_meta_val_str(model, "general.parameter_count", param_size_buf, sizeof(param_size_buf));

json models = {
{"models", {
{
{"name", params.model_alias.empty() ? params.model.path : params.model_alias},
{"model", params.model_alias.empty() ? params.model.path : params.model_alias},
{"modified_at", modified_buf},
{"size", file_stat.st_size},
{"digest", ""}, // TODO: add digest
{"type", "model"},
{"description", ""},
{"tags", {arch_buf}},
{"capabilities", {"completion"}},
{"parameters", ""}, // TODO: add parameters
{"details", {
{"parent_model", ""}, // TODO: Add parent_model
{"format", "gguf"},
{"family", arch_buf},
{"families", {arch_buf}},
{"parameter_size", param_size_buf},
{"quantization_level", ""} // TODO: add quantization level if available
}}
}
}},
{"object", "list"},
{"data", {
{
Expand All @@ -4420,7 +4483,7 @@ int main(int argc, char ** argv) {
{"owned_by", "llamacpp"},
{"meta", model_meta},
},
}}
}}
};

res_ok(res, models);
Expand Down Expand Up @@ -4748,11 +4811,13 @@ int main(int argc, char ** argv) {
svr->Post("/api/show", handle_api_show);
svr->Get ("/models", handle_models); // public endpoint (no API key check)
svr->Get ("/v1/models", handle_models); // public endpoint (no API key check)
svr->Get ("/api/tags", handle_models); // ollama specific endpoint. public endpoint (no API key check)
svr->Post("/completion", handle_completions); // legacy
svr->Post("/completions", handle_completions);
svr->Post("/v1/completions", handle_completions_oai);
svr->Post("/chat/completions", handle_chat_completions);
svr->Post("/v1/chat/completions", handle_chat_completions);
svr->Post("/api/chat", handle_chat_completions); // ollama specific endpoint
svr->Post("/infill", handle_infill);
svr->Post("/embedding", handle_embeddings); // legacy
svr->Post("/embeddings", handle_embeddings);
Expand Down
Loading