Distribute and run LLMs with a single file.
C++ cross-platformggufllama-cpplocal-ailocal-inferencelocal-llmopen-source-aisingle-file-executablespeech-to-text
diff --git a/tools/server/server-http.h b/tools/server/server-http.h
--- a/llama.cpp/tools/server/server-http.h
+++ b/llama.cpp/tools/server/server-http.h
@@ -91,11 +91,16 @@ struct server_http_context {
std::string path_prefix;
int port = 9931;
bool is_ssl = false;
+ bool is_bound = false;
server_http_context();
~server_http_context();
bool init(const common_params & params);
+ // creates and binds the listening sockets; start() calls it if needed.
+ // llamafile binds early so the sockets exist before the sandbox is
+ // installed (a Unix domain socket can't be created under pledge "anet")
+ bool bind_socket();
bool start();
void stop() const;
void join();