Command-line interface for the Elnora AI Platform.
Elnora is an AI agent for biotech and pharma teams that generates, optimizes, and troubleshoots preclinical lab protocols. She learns from both successful and failed experiments, building knowledge that compounds with every project.
The CLI gives you full access to the platform from your terminal — collaborate with the agent, manage protocols and files, and integrate Elnora into your automation pipelines.
# macOS / Linux
curl -fsSL https://cli.elnora.ai/install.sh | bash
# Windows (PowerShell)
irm https://cli.elnora.ai/install.ps1 | iex
# npm (requires Node.js 20+)
npm install -g @elnora-ai/cli
# Homebrew
brew install elnora-ai/cli/elnora
After installing the CLI, configure your AI coding tools:
elnora setup
This auto-detects which tools you have installed and configures them:
elnora-plugins marketplace and enables 9 skills (elnora-tasks, elnora-projects, elnora-files, etc.) so you can use natural language.https://mcp.elnora.ai/mcp with your API key.Target a specific tool: elnora setup claude · elnora setup cursor · elnora setup vscode · elnora setup codex.
Once configured, restart Claude Code and try any natural-language request:
“Use Elnora to generate a PCR protocol for BRCA1 exon 11” “Use Elnora to list my projects” “Use Elnora to optimize the protocol in task abc-123”
Claude will invoke the elnora CLI under the hood using the bundled skills. No need to memorize command syntax.
| What | Where |
|---|---|
| CLI binary (macOS/Linux) | ~/.local/bin/elnora |
| CLI binary (Windows) | %USERPROFILE%\.elnora\bin\elnora.exe |
| Auth / profiles | ~/.elnora/profiles.toml (mode 0600) |
| Claude Code skills | ~/.claude/plugins/marketplaces/elnora-plugins/elnora/skills/ (auto-updated) |
| Claude Code settings | ~/.claude/settings.json — look for elnora@elnora-plugins: true |
If elnora setup claude doesn’t work for some reason, you can install the plugin manually:
/pluginElnora-AI/elnora-plugins/plugin again → Plugins → Enable elnoraelnora auth login
elnora projects list
elnora tasks create --project <id> --message "Analyze kinase inhibitors"
elnora tasks send <task-id> --message "Focus on selectivity" --stream
| Group | Commands |
|---|---|
auth |
login, logout, status, profiles, validate |
projects |
list, get, create, update, archive, members, add-member, update-role, remove-member, leave |
tasks |
list, get, create, send (–stream, –wait), messages, update, archive |
files |
list, get, content, create, upload, download, update, archive, versions, fork, promote, working-copy, commit |
orgs |
list, get, create, update, members, billing, invite, invitations, resend-invite, cancel-invite, delete |
folders |
list, create, rename, move, delete |
search |
tasks, files, all, file-content |
library |
files, folders, create-folder, rename-folder, delete-folder |
account |
get, update, agreements, accept-terms, delete, users |
api-keys |
create, list, revoke, get-policy, set-policy |
audit |
list |
feedback |
submit |
flags |
list, get, set |
health |
check |
elnora doctor # Check API, auth, version, config
elnora whoami # Show current profile and org
elnora open [page] # Open platform in browser (docs, keys, billing)
elnora completion <shell> # Generate shell completions (bash, zsh, fish)
elnora mcp serve # Start built-in MCP server
Real-time agent responses directly in your terminal:
# Stream agent output as it generates
elnora tasks send <task-id> --message "Optimize this protocol" --stream
# Wait for the complete response (polling)
elnora tasks send <task-id> --message "Optimize this protocol" --wait
Status output (thinking, tool calls) goes to stderr. Content goes to stdout, so streaming is pipeable:
elnora tasks send <task-id> --message "Analyze this" --stream > response.txt
The CLI includes a built-in MCP server that exposes all commands as tools:
# Start HTTP MCP server (for deployment)
elnora mcp serve --http
# Start stdio MCP server (for local MCP clients)
elnora mcp serve --stdio
Configure in Claude Code or Cursor:
{
"elnora": {
"type": "http",
"url": "https://mcp.elnora.ai/mcp"
}
}
# Login with API key
elnora auth login
# Multi-org profiles
elnora auth login --profile university
elnora --profile university projects list
# Check status
elnora doctor
API keys are stored in ~/.elnora/profiles.toml with secure file permissions (0600).
elnora projects list # JSON (default)
elnora projects list --compact # Compact JSON
elnora projects list --output csv # CSV
elnora projects list --fields id,name # Field filtering
elnora projects list --json # Force JSON (even in TTY)
Apache-2.0