CLI for local agents like Claude Code and Cowork. MCP server for web platforms like claude.ai, Manus, and ChatGPT. One product, two paths — same 20+ commands.
Ask your Claude Code, Cowork, or other local agent to install the CLI and login for you.
One MCP tool exposes the entire DeepCell CLI to web-based AI agents. Enterprise-grade OAuth 2.1 security with PKCE.
# Install the CLI
$ pip install deepcell-cli
# Ensure the CLI is on your PATH
$ export PATH="$HOME/.local/bin:$PATH"
# Authenticate
$ deepcell loginTo make the PATH change permanent, add export PATH="$HOME/.local/bin:$PATH" to your ~/.zshrc.
Add this URL as a remote MCP server in Claude.ai settings:
https://your-domain.com/mcpGo to Claude.ai → Settings → Integrations → Add MCP Server → paste the URL above. OAuth authentication is handled automatically.
Start a chat in Claude Code
“Use deepcell cli to create a dcf model, starting with deepcell --help”
Start a chat in your platform
“Use deepcell tool to create a dcf model”
Works with
Claude Code
Anthropic's CLI agent for coding and terminal workflows.
Cowork
Anthropic's agentic environment for teams and enterprises.
OpenWork
Open-source desktop AI coworker with MCP and multi-LLM support.
Works with
claude.ai
Anthropic's web-based AI assistant with native MCP support.
Manus
Autonomous AI agent platform with MCP tool integration.
Any MCP client
Any platform supporting the Model Context Protocol standard.
Ask your Claude Code, Cowork, or other local agent to install the CLI and login for you.
# Install the CLI
$ pip install deepcell-cli
# Ensure the CLI is on your PATH
$ export PATH="$HOME/.local/bin:$PATH"
# Authenticate
$ deepcell loginTo make the PATH change permanent, add export PATH="$HOME/.local/bin:$PATH" to your ~/.zshrc.
Start a chat in Claude Code
“Use deepcell cli to create a dcf model, starting with deepcell --help”
Works with
Claude Code
Anthropic's CLI agent for coding and terminal workflows.
Cowork
Anthropic's agentic environment for teams and enterprises.
OpenWork
Open-source desktop AI coworker with MCP and multi-LLM support.
Every command works the same whether invoked via CLI or MCP — structured JSON output, full audit trail.
Every command returns structured JSON. Agents parse directly — no scraping.
Git-like operations — log, diff, restore — for structured financial data.
Edit single values or pass a JSON file for bulk updates across models.
Export to .xlsx with live formulas intact or values-only for downstream use.
Address any value by Item, Context, and Status — not row and column.
Industry-standard authorization. No client secrets required for web agents.
15-minute JWT lifetime with refresh tokens rotated on every use.
Blocked commands prevent auth-bypass. Every MCP call is logged for audit.
Every command returns plain text — token-light, easy for any agent to parse and act on.
# Query a projected revenue value
$ deepcell query model.deepcell Revenue FY2025E projected
1500000
# Batch edit from a JSON file
$ deepcell edit model.deepcell --batch changes.json
3 edit(s) applied rev:a1b2c3d4
✓ 3 edit(s) applied
# Export to Excel with live formulas
$ deepcell to-excel model.deepcell -o forecast.xlsx --formulas
✓ Exported to forecast.xlsx (28473 bytes)app.product.connect.steps.0.desc
app.product.connect.steps.1.desc
app.product.connect.steps.2.desc
Get started
Install the CLI or connect via MCP — either way, your agent gets 20+ commands for querying, editing, and exporting financial models.