{"id":1980,"date":"2026-07-18T06:32:05","date_gmt":"2026-07-18T13:32:05","guid":{"rendered":"http:\/\/macdaddy4sure.ai\/?p=1980"},"modified":"2026-07-18T06:32:05","modified_gmt":"2026-07-18T13:32:05","slug":"environment-variables-for-_augmentedintelligence","status":"publish","type":"post","link":"http:\/\/macdaddy4sure.ai\/index.php\/2026\/07\/18\/environment-variables-for-_augmentedintelligence\/","title":{"rendered":"Environment variables for _AugmentedIntelligence"},"content":{"rendered":"\n

Recommended defaults for a dev workstation<\/strong>. Empty string means \u201cunset \/ feature off until configured.\u201d<\/p>\n\n\n\n

Also see: RemoteCUDA.md (RemoteCUDA.md) \u00b7 BuildConfigurations.md (BuildConfigurations.md) \u00b7 PluginBridges.md (PluginBridges.md)<\/p>\n\n\n\n

Quick start (dev defaults)<\/h2>\n\n\n\n

[bat]
<\/em>REM — Core (usually leave unset; Settings\/MySQL file wins) —
REM set AI_MYSQL_HOST=127.0.0.1
REM set AI_MYSQL_USER=root
REM set AI_MYSQL_PASSWORD=
REM set AI_MYSQL_DATABASE=cognitive_loop

REM — Remote CUDA worker (optional GPU box) —
set AI_REMOTE_CUDA_ENABLED=1
set AI_REMOTE_CUDA_HOST=127.0.0.1
set AI_REMOTE_CUDA_PORT=7720
set AI_REMOTE_CUDA_TOKEN=dev-token
set AI_REMOTE_CUDA_TIMEOUT_MS=15000

REM On the GPU machine:
REM   set AI_REMOTE_CUDA_TOKEN=dev-token
REM   set AI_REMOTE_CUDA_PORT=7720
REM   python tools\\remote_cuda_worker.py<\/p>\n\n\n\n

Remote CUDA \/ HardwareAcceleration<\/h2>\n\n\n\n
Variable<\/strong><\/td>Default<\/strong><\/td>Meaning<\/strong><\/td><\/tr>
`AI_REMOTE_CUDA_ENABLED`<\/strong><\/td>1<\/td>Master enable for _RemoteCuda client (0\/false\/off disables)<\/td><\/tr>
`AI_REMOTE_CUDA_HOST`<\/strong><\/td>(empty)<\/em><\/td>Single worker host: host, host:port, or http:\/\/host:port<\/td><\/tr>
`AI_REMOTE_CUDA_HOSTS`<\/strong><\/td>(empty)<\/em><\/td>Comma\/space-separated worker list (pool; least-loaded pick)<\/td><\/tr>
`AI_REMOTE_CUDA_PORT`<\/strong><\/td>7720<\/td>Default port when host has no :port<\/td><\/tr>
`AI_REMOTE_CUDA_TOKEN`<\/strong><\/td>(empty)<\/em><\/td>Bearer \/ JSON token for worker auth<\/td><\/tr>
`AI_REMOTE_CUDA_TIMEOUT_MS`<\/strong><\/td>15000<\/td>HTTP timeout for remote RPC<\/td><\/tr>
`AI_REMOTE_CUDA_BIND`<\/strong><\/td>0.0.0.0<\/td>Worker listen address (remote_cuda_worker.py)<\/td><\/tr>
`AI_CUDA_REMOTE_HOST`<\/strong><\/td>(empty)<\/em><\/td>Alias for AI_REMOTE_CUDA_HOST<\/td><\/tr>
`AI_HARDWARE_REMOTE_TOKEN`<\/strong><\/td>(empty)<\/em><\/td>Fallback token if AI_REMOTE_CUDA_TOKEN unset<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

In-app Settings globals<\/strong> (not env, but related):<\/p>\n\n\n\n

cuda_accelleration_remote_enabled, cuda_accelleration_remote_hostname,<\/p>\n\n\n\n

tensor_accelleration_remote_*, hardware_remote_rpc_port (default 7720<\/strong>),<\/p>\n\n\n\n

hardware_remote_rpc_timeout_ms (15000<\/strong>), hardware_remote_rpc_token.<\/p>\n\n\n\n

Worker:<\/strong> python tools\/remote_cuda_worker.py<\/p>\n\n\n\n

Client commands:<\/strong> cuda remote status|ping|hello|bench|set|job|tensor|checkpoint|\u2026<\/p>\n\n\n\n

Database \/ MySQL<\/h2>\n\n\n\n
Variable<\/strong><\/td>Default<\/strong><\/td>Meaning<\/strong><\/td><\/tr>
MySQL connection<\/td>from Settings \/ AD Credential Manager<\/em><\/td>Host app uses _Settings + optional CLI; env override patterns vary by build<\/td><\/tr>
Common convention<\/td>127.0.0.1 \/ root \/ empty password \/ cognitive_loop<\/td>Local XAMPP\/WAMP style lab<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

If you use env-based secrets tooling, prefer OS Credential Manager \/ GPO (_Settings::ApplyAdPasswords) over plain env in production.<\/p>\n\n\n\n

LLM \/ remote inference (non-CUDA)<\/h2>\n\n\n\n
Variable<\/strong><\/td>Default<\/strong><\/td>Meaning<\/strong><\/td><\/tr>
Provider API keys<\/td>unset<\/em><\/td>OpenAI\/Grok\/Gemini\/etc. usually via Settings UI \/ encrypted store, not required env<\/td><\/tr>
Ollama \/ remote LLM<\/td>Settings<\/em><\/td>remote_llm_server_enabled + hostname in settings<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Bridges \/ games<\/h2>\n\n\n\n
Variable<\/strong><\/td>Default<\/strong><\/td>Meaning<\/strong><\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
Bridge tokens<\/td>configured in-app<\/em><\/td>Minecraft\/SE\/GameMod tokens live in bridge config files \/ commands (minecraft bridge configure \u2026)<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
Lab hosts<\/td>e.g. 10.0.0.151<\/td>Documented in Minecraft\/SE docs for EARTH\/JUPITER<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_CREO_PROFILE<\/td>configurator<\/td>Creo strategy profile: live \\<\/td>batch \\<\/td>configurator \\<\/td>file \\<\/td>toolkit<\/td><\/tr>
AI_CREO_API<\/td>mock<\/td>Creo worker API: mock \\<\/td>vb \\<\/td>jlink \\<\/td>trail \\<\/td>toolkit<\/td><\/tr>
AI_CREO_HOST<\/td>127.0.0.1<\/td>Creo worker host<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_CREO_BRIDGE_PORT<\/td>8777<\/td>Creo worker TCP port<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_CREO_TOKEN<\/td>change-me<\/td>Shared bridge token<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_CREO_WORKDIR<\/td>creo\/work<\/td>Sandbox for exports \/ jobs<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_CREO_WORKER<\/td>(unset)<\/em><\/td>Path to worker script\/exe (tools\/creo_worker\/creo_worker.py)<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_CREO_EXE<\/td>(unset)<\/em><\/td>Path to Creo Parametric executable<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_CREO_DRY_RUN<\/td>on<\/td>Default dry-run for Creo exports<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_BLENDER_PROFILE<\/td>template<\/td>Blender profile: interactive \\<\/td>headless \\<\/td>template \\<\/td>file \\<\/td>advanced<\/td><\/tr>
AI_BLENDER_API<\/td>mock<\/td>mock or live worker<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_BLENDER_HOST \/ AI_BLENDER_PORT<\/td>127.0.0.1 \/ 8775<\/td>Blender worker TCP<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_BLENDER_TOKEN<\/td>change-me<\/td>Blender bridge token<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_BLENDER_WORKDIR<\/td>blender\/work<\/td>Sandbox<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_BLENDER_WORKER \/ AI_BLENDER_EXE<\/td>(unset)<\/em><\/td>Worker script \/ blender.exe<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_BLENDER_DRY_RUN<\/td>on<\/td>Default dry-run<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_C4D_PROFILE<\/td>template<\/td>C4D profile: live \\<\/td>batch \\<\/td>template \\<\/td>file \\<\/td>sdk<\/td><\/tr>
AI_C4D_API<\/td>mock<\/td>mock or live worker<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_C4D_HOST \/ AI_C4D_PORT<\/td>127.0.0.1 \/ 8776<\/td>C4D worker TCP<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_C4D_TOKEN<\/td>change-me<\/td>C4D bridge token<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_C4D_WORKDIR<\/td>c4d\/work<\/td>Sandbox<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_C4D_WORKER \/ AI_C4D_EXE<\/td>(unset)<\/em><\/td>Worker \/ Cinema 4D exe<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_C4D_DRY_RUN<\/td>on<\/td>Default dry-run<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_WEATHER_LAT \/ AI_WEATHER_LON<\/td>Seattle defaults<\/td>Weather\/storm point<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_WEATHER_LOCATION<\/td>Seattle<\/td>Location label<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_WEATHER_NWS_AREA<\/td>(unset)<\/em><\/td>Optional US state (e.g. WA) for NWS alerts<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_WEATHER_MOCK \/ AI_WEATHER_FORCE_MOCK<\/td>off<\/td>Use fixtures (offline)<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_WEATHER_USER_AGENT<\/td>AugmentedIntelligence\/\u2026<\/td>Required style header for NWS<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_WEATHER_TIMEOUT<\/td>20<\/td>HTTP timeout seconds<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_NEMO_URL<\/td>http:\/\/127.0.0.1:8000<\/td>NeMo Guardrails \/ NIM base URL<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_NEMO_TOKEN<\/td>(unset)<\/em><\/td>Optional bearer<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_NEMO_PROFILE<\/td>mock<\/td>mock \\<\/td>guardrails \\<\/td>nim \\<\/td>speech \\<\/td>\u2026<\/td><\/tr>
AI_NEMO_DRY_RUN<\/td>on<\/td>Prefer local rails when not live<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_TF_PYTHON<\/td>python<\/td>Python for TF pipelines<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_TF_TRAIN_ROOT<\/td>tensorflow_training<\/td>Pipeline root<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
AI_TF_TRAIN_PROFILE<\/td>mock<\/td>mock \\<\/td>pipeline \\<\/td>finetune \\<\/td>native<\/td> <\/td><\/tr>
AI_TF_TRAIN_DRY_RUN<\/td>on<\/td>Default dry-run for training jobs<\/td> <\/td> <\/td> <\/td> <\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

No required env for bridges; dry-run on<\/strong> by default in BridgeManager. Creo\/Blender\/C4D work offline with *_API=mock (default). Weather uses Open-Meteo + NWS + USGS \u2014 see WeatherHazards.md (WeatherHazards.md).<\/p>\n\n\n\n

Plugins<\/h2>\n\n\n\n
Variable<\/strong><\/td>Default<\/strong><\/td>Meaning<\/strong><\/td><\/tr>
(none required)<\/em><\/td>\u2014<\/td>Plugins load in-process; dynamic DLL via plugins load<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Sandbox paths: plugins\/data\/ (HostApi FS).<\/p>\n\n\n\n

CUDA \/ GPU local<\/h2>\n\n\n\n
Variable<\/strong><\/td>Default<\/strong><\/td>Meaning<\/strong><\/td><\/tr>
`CUDA_VISIBLE_DEVICES`<\/strong><\/td>(driver default)<\/em><\/td>Standard NVIDIA: which GPUs are visible<\/td><\/tr>
`CUDA_PATH`<\/strong><\/td>toolkit install<\/em><\/td>MSVC\/CUDA toolkit discovery (system install)<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Build-time: USE_CUDA \/ AI_CPU_ONLY in VS configurations (not runtime env).<\/p>\n\n\n\n

Networking \/ portal<\/h2>\n\n\n\n
Variable<\/strong><\/td>Default<\/strong><\/td>Meaning<\/strong><\/td><\/tr>
Portal \/ web<\/td>Settings<\/em><\/td>program_state_portal_*, web server ports in settings<\/td><\/tr>
Hardware RPC port<\/td>7720<\/td>hardware_remote_rpc_port<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Testing \/ CI<\/h2>\n\n\n\n
Variable<\/strong><\/td>Default<\/strong><\/td>Meaning<\/strong><\/td><\/tr>
`AI_EXE`<\/strong><\/td>(auto-discover)<\/em><\/td>Full path to _AugmentedIntelligence.exe for tests\/run_command_batch.ps1<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Recommended profiles<\/h2>\n\n\n\n

A. Laptop CPU-only<\/h3>\n\n\n\n

[bat]
<\/em>set AI_REMOTE_CUDA_ENABLED=0
REM do not set AI_REMOTE_CUDA_HOST<\/p>\n\n\n\n

B. Local GPU + local worker (same machine)<\/h3>\n\n\n\n

[bat]
<\/em>set AI_REMOTE_CUDA_ENABLED=1
set AI_REMOTE_CUDA_HOST=127.0.0.1
set AI_REMOTE_CUDA_PORT=7720
set AI_REMOTE_CUDA_TOKEN=dev-token<\/p>\n\n\n\n

C. Workstation + remote GPU server<\/h3>\n\n\n\n

[bat]
<\/em>set AI_REMOTE_CUDA_ENABLED=1
set AI_REMOTE_CUDA_HOSTS=http:\/\/10.0.0.50:7720,http:\/\/10.0.0.51:7720
set AI_REMOTE_CUDA_TOKEN=prod-long-random-token
set AI_REMOTE_CUDA_TIMEOUT_MS=30000<\/p>\n\n\n\n

D. Match HardwareAcceleration remote Tensor path<\/h3>\n\n\n\n

[bat]
<\/em>set AI_REMOTE_CUDA_HOST=10.0.0.50
set AI_REMOTE_CUDA_PORT=7720
set AI_REMOTE_CUDA_TOKEN=same-as-hardware_remote_rpc_token<\/p>\n\n\n\n

Also enable remote CUDA\/tensor hostnames in Settings if you use the older HA code paths.<\/p>\n\n\n\n

Security notes<\/h2>\n\n\n\n
    \n
  1. Never commit real tokens.<\/strong> Use dev-token only on localhost.\u00a0<\/li>\n\n\n\n
  2. Bind workers to LAN or localhost; use firewall + strong token in production.\u00a0<\/li>\n\n\n\n
  3. Kill-switch \/ safe mode cancels<\/strong> remote CUDA jobs.\u00a0<\/li>\n\n\n\n
  4. Remote matmul always falls back to local GEMM<\/strong> if the worker is down.<\/li>\n<\/ol>\n\n\n\n

    Verification<\/h2>\n\n\n\n

    [text]
    <\/em>cuda remote refresh
    cuda remote status
    cuda remote ping
    cuda remote bench 128
    nn device list
    nn device set global train remote<\/p>\n\n\n\n

    Worker logs should show \/v1\/ping and \/infer or \/v1\/matmul hits.<\/p>\n","protected":false},"excerpt":{"rendered":"

    Recommended defaults for a dev workstation. Empty string means \u201cunset \/ feature off until configured.\u201d Also see: RemoteCUDA.md (RemoteCUDA.md) \u00b7 BuildConfigurations.md (BuildConfigurations.md) \u00b7 PluginBridges.md (PluginBridges.md) Quick start (dev defaults) [bat]REM — Core (usually leave unset; Settings\/MySQL file wins) —REM set AI_MYSQL_HOST=127.0.0.1REM set AI_MYSQL_USER=rootREM set AI_MYSQL_PASSWORD=REM set AI_MYSQL_DATABASE=cognitive_loop REM — Remote CUDA worker (optional GPU box) […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-1980","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/posts\/1980","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/comments?post=1980"}],"version-history":[{"count":1,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/posts\/1980\/revisions"}],"predecessor-version":[{"id":1981,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/posts\/1980\/revisions\/1981"}],"wp:attachment":[{"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/media?parent=1980"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/categories?post=1980"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/tags?post=1980"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}