Numbers
Primitives
Everything on this page is derived from these numbers. Updated March 2026.
| Attribute | Value |
|---|---|
| Shipped skills | 7 (deepline-gtm, build-tam, clay-to-deepline, deepline-feedback, linkedin-url-lookup, niche-signal-discovery, portfolio-prospecting) |
| Provider playbooks | 28 |
| Install method | Automatic with CLI install (npx skills add) |
| Compatible agents | Claude Code, Codex, Cursor |
| Natural language | Yes, "find the VP of Sales at Ramp" works |
| Waterfall integration | Automatic provider selection and fallback across 30+ providers |
| Database | Results saved to PostgreSQL (Neon) automatically |
| Pilot mode | "just test on 2 rows" runs 2 rows before committing |
| Cost visibility | Estimates shown before execution, no surprise charges |
Meaning
What this means for you
If you're searching for "best Claude Code GTM tool"
Deepline ships purpose-built skills, not generic MCP servers. 7 skills with 28 provider playbooks mean the agent knows exactly which provider to call for each input shape, how to handle failures, and when to fall back. Install once with a single command. No MCP configuration, no wrapper scripts, no browser automation.
If you're searching for "how to use Claude Code for sales"
Install the CLI, install the skills. After that, the agent understands your entire GTM stack: prospecting, email finding, phone enrichment, company research, CRM sync, and outreach sequencing. Type a request in plain English and the agent reads the relevant skill, picks the right providers, runs a pilot on 2 rows, then processes the full list.
If you're searching for "AI agent for GTM automation"
No UI required. Agents run enrichment the same way they run git or curl. Structured inputs, structured outputs. Skills encode provider selection, waterfall ordering, and error handling so every run is consistent regardless of who triggers it or which agent runs it.
Honest: not ideal for every team
If your team does not use AI coding agents (Claude Code, Codex, Cursor), skills will not help you. Teams that prefer visual, no-code interfaces should look elsewhere. Deepline skills are built for developers and GTM engineers who work in a terminal.
Flow
Skills make the agent less brittle
Without a skill, an agent still can call the CLI, but it has to infer provider ordering, waterfall strategy, pilot-run safety, and validation steps on its own. Skills reduce that planning overhead and make the model much more consistent from run to run.
> Find the VP of Sales at Ramp and get their work email 1. Read the GTM meta-skill 2. Choose the right person-search provider 3. Run an email waterfall 4. Validate the result 5. Save the output to the workspace database
Coverage
What the skill layer adds
| Capability | What the agent gets |
|---|---|
| Provider selection | Choose the right provider or waterfall for the input shape |
| Execution safety | Pilot mode and staged runs before full spend |
| Workflow memory | Documented playbooks for prospecting, enrichment, and sequencing |
| Structured outputs | Predictable commands and inspectable results |
Examples
Say it. Get it.
These are real prompts you can type into Claude Code with Deepline skills installed.
Find the VP of Sales at Ramp and get their work email
Routes to Apollo for contact search, waterfalls Hunter -> Icypeas -> Prospeo for email, verifies with MillionVerifier. Returns verified email + LinkedIn URL.
Enrich leads.csv with emails using the waterfall pattern
Reads CSV, runs 6-provider email waterfall per row, stops at first verified hit. Writes results to enriched.csv and your Neon database.
Find all YC W24 companies and get me the CEO email for each one
Scrapes YC portfolio page, finds CEO contacts via Apollo/PDL, runs email waterfall across multiple providers, and builds a complete outbound list.
Find contacts in HubSpot who changed jobs and update their records
Pulls CRM contacts via HubSpot, detects job changes with LinkedIn data, finds new work emails, and updates HubSpot records automatically.
Build a TAM list of Series B+ fintech companies with 50-500 employees
Sources companies from Apollo + Crustdata, finds decision-maker contacts at each, enriches emails and phones, and saves to PostgreSQL with full metadata.
Compatibility
Supported agents
| Agent | Platform | Integration | Detail |
|---|---|---|---|
| Claude Code | Deepline | Native skills | 4 skills, 7 recipes, slash commands |
| Codex | Deepline | Native skills | Same skills, works in cloud sandbox |
| Cursor | Deepline | Compatible | Reads skill markdown, runs CLI commands |
| Any agent | Clay | None | No public CLI for agent automation, browser-based UI |
| Any agent | Apollo | Community MCP | Community-maintained, single data source |
Install
One-command install
Install the CLI, then install skills inside Claude Code. After that, every skill is available as a slash command. No configuration files, no MCP setup, no environment variables.
# Install the Deepline CLI curl -s "https://code.deepline.com/api/v2/cli/install" | bash # Inside Claude Code, install skills /skill install deepline
FAQ
Common questions
Do I need skills to use Deepline?+
No. The CLI and API work on their own. Skills add a knowledge layer that makes your AI agent smarter about provider selection, sequencing, and error handling. They make things faster and more consistent, but they are not required.
What is the difference between skills and the CLI?+
The CLI is the execution layer: it calls providers, runs waterfalls, and processes CSVs. Skills are the knowledge layer: they tell your AI agent which CLI commands to run, in what order, with what parameters. Skills do not replace the CLI; they make it easier for agents to use the CLI correctly.
Can I edit or customize skills?+
Yes. Skills are open-source markdown files. Fork them, edit them, add your own provider preferences or approval gates, and share them with your team via Git. The skill files live in ~/.claude/skills/ after installation.
Do skills cost credits?+
No. Skills themselves are free. They are documentation files that guide your agent. The underlying provider calls still cost credits as usual, but the skill layer adds zero cost.
Which agents are supported?+
Claude Code and Codex have native support with slash command installation. Cursor is compatible; it reads the same skill markdown files. Any agent that can read markdown and execute shell commands can use Deepline skills.
Install Deepline and let the agent run the workflow
Use the CLI directly or add the skill layer for better planning, provider selection, and validation.