CroweLM Deepcore

A focused command-line interface for the CroweLM Deepcore model, served via Crowe Logic.

Install

uv venv
uv pip install -p .venv -e .

Configure

Set the backend env vars (a .env file in the working directory is loaded automatically):

# default backend: azure
AZURE_CORE_ENDPOINT=https://<resource>.openai.azure.com
AZURE_CORE_API_KEY=<key>
# optional overrides
DEEPCORE_API_VERSION=2024-08-01-preview

# or route through the Crowe Logic Foundry control plane:
DEEPCORE_BACKEND=foundry
FOUNDRY_BASE_URL=https://<control-plane>
FOUNDRY_API_KEY=<token>

Use

deepcore                 # interactive agent chat
deepcore run "question"  # one-shot, pipe-friendly (answer on stdout)
deepcore tools           # list the agent's tools
deepcore info            # model + backend status
deepcore doctor          # diagnose config + reachability
deepcore run --no-tools "..."   # plain chat, no tools
deepcore run --yes "..."        # auto-approve tool actions

Fusion (stacking models for stronger output)

deepcore can stack the answerer with a separate reasoning model. All modes compose hosted backends (no GPU/weight-merging), so they are API-call stacking.

deepcore run --fuse reason "hard question"     # reasoner thinks, answerer answers
deepcore run --fuse escalate "question"        # answer first; escalate to reasoner only if unsure
deepcore run --deep "question"                 # heavy: many models in parallel + a judge (slow)
deepcore run --dual "A,B" "question"           # compare two deployments side by side
deepcore run --dual "A,B" --synth "question"   # ...and synthesize a merged answer

--fuse also works in interactive chat. Set a default with DEEPCORE_FUSION=reason.

Agent and tools

deepcore is an agent: it can call tools to inspect and change your project.

• Read-only (run automatically): read_file, list_dir, glob, grep, ast_symbols, ast_show_symbol, web_fetch, web_search, analyze_image.
• Mutating / executing (require confirmation): write_file, edit_file, ast_replace_symbol, run_shell, run_code.

web_search needs DEEPCORE_SEARCH_API_KEY; analyze_image needs DEEPCORE_VISION_DEPLOYMENT (a multimodal model). Both degrade gracefully when unconfigured.

In interactive chat, mutating actions prompt for y/n. In run (non-interactive) they are denied unless you pass --yes. ast_* tools are multi-language via tree-sitter; run_code executes in a Docker sandbox when available, else a timeboxed local subprocess.

Configuration reference

CroweLM Deepcore is served via Crowe Logic infrastructure. https://crowelogic.com