A solo founder can run a working outbound motion with three paid tools: Claude Code to research and write, Databar for company data, contacts and verified emails, and a sending tool for delivery. At September 2026 list prices that is roughly $166 to $178 a month before domains. The real work is setting it up once and keeping a weekly routine.

This guide covers what the stack costs, how to set it up in an afternoon, the weekly routine that keeps it running, and the guardrails that stop an agent from spending your credits on the wrong list. For the full step-by-step campaign build, see how to build an outbound campaign with Claude Code. This post is about running that system alone, week after week.

What does a one-person GTM stack actually need to do?

Strip outbound down and there are six jobs. A team splits them across people. Alone, you need each one covered by a tool or a habit:

1. Decide who to target. Your ICP, written down, not held in your head.

2. Find accounts. Companies that match, from a search or a list you already have.

3. Find the right person and a verified email. The part that burns the most money when done badly.

4. Write the message. Short, specific, based on something true about the account.

5. Send without wrecking your domain. Separate sending domains, warm-up, volume limits.

6. Learn from replies. Record what worked so the next batch is better.

Claude Code covers jobs 1, 4 and 6 through project files it reads every session. Databar covers 2 and 3. The sending tool covers 5.

How much does a solo founder GTM stack cost per month?

Prices below were checked on each vendor's pricing page in September 2026.

Claude Pro plus Databar Build plus one sending tool comes to $166 or $178 a month. Add a few secondary domains for sending, and move to Claude Max only if you regularly hit Pro's usage limits, which Claude Code shares with your chats.

For comparison, Clay's Launch plan was listed at $185/month billed monthly on the same day, with 2,500 data credits and 15,000 actions. Clay also has a free plan with 100 data credits a month. It is a strong product if you like building in its tables; the stack above is simply cheaper to start and built around working from Claude Code.

What will 5,000 Databar credits cover?

Every enrichment has its own credit price, and only lookups that return data are charged. Here is an illustration using catalog prices from September 2026, assuming the first provider in each step returns data:

• Company basics (headcount and funding) from the "Lookup company data" waterfall starting with Muraena: 2 credits.

• Decision-maker by title with Explee's "Find people by website and job title": 2 credits.

• Work email from the email waterfall starting with Icypeas: 3 credits, plus 1 credit for verification.

That is about 8 credits per fully worked lead in the best case, so 5,000 credits stretch to several hundred leads a month. When the first provider misses, the next one in the chain may cost more, so budget for a higher average and check your balance after the first batch. Unused credits on monthly plans roll over for one month.

How do you set it up in one afternoon?

1. Create a project folder and a short CLAUDE.md

Claude Code reads CLAUDE.md from your project folder at the start of every session. Keep it to what you would otherwise re-explain every time:

# Outbound for Acme

## Product
One paragraph: what it does, who it helps, the problem it removes.

## ICP
- Titles: Head of RevOps, VP Sales
- Companies: B2B SaaS, 30-200 employees, US and UK
- Exclude: agencies, current customers (see customers.csv), competitors

## Voice
- Under 90 words. One question. No "hope this finds you well".

## Rules
- Show a cost estimate and wait for my OK before any enrichment over 50 rows.
- Never send email. Output a CSV for me to review.

## Learnings
- (add after each batch: what got replies, what didn't)

# Outbound for Acme

## Product
One paragraph: what it does, who it helps, the problem it removes.

## ICP
- Titles: Head of RevOps, VP Sales
- Companies: B2B SaaS, 30-200 employees, US and UK
- Exclude: agencies, current customers (see customers.csv), competitors

## Voice
- Under 90 words. One question. No "hope this finds you well".

## Rules
- Show a cost estimate and wait for my OK before any enrichment over 50 rows.
- Never send email. Output a CSV for me to review.

## Learnings
- (add after each batch: what got replies, what didn't)

# Outbound for Acme

## Product
One paragraph: what it does, who it helps, the problem it removes.

## ICP
- Titles: Head of RevOps, VP Sales
- Companies: B2B SaaS, 30-200 employees, US and UK
- Exclude: agencies, current customers (see customers.csv), competitors

## Voice
- Under 90 words. One question. No "hope this finds you well".

## Rules
- Show a cost estimate and wait for my OK before any enrichment over 50 rows.
- Never send email. Output a CSV for me to review.

## Learnings
- (add after each batch: what got replies, what didn't)

# Outbound for Acme

## Product
One paragraph: what it does, who it helps, the problem it removes.

## ICP
- Titles: Head of RevOps, VP Sales
- Companies: B2B SaaS, 30-200 employees, US and UK
- Exclude: agencies, current customers (see customers.csv), competitors

## Voice
- Under 90 words. One question. No "hope this finds you well".

## Rules
- Show a cost estimate and wait for my OK before any enrichment over 50 rows.
- Never send email. Output a CSV for me to review.

## Learnings
- (add after each batch: what got replies, what didn't)

The CLAUDE.md template for GTM teams has a longer version if you want more structure.

2. Connect Databar to Claude Code

Grab your API key from your Databar workspace under Integrations, then run:

claude mcp add --transport http databar https://mcp.databar.ai/mcp \
  --header "Authorization: Bearer YOUR_DATABAR_API_KEY"

claude mcp add --transport http databar https://mcp.databar.ai/mcp \
  --header "Authorization: Bearer YOUR_DATABAR_API_KEY"

claude mcp add --transport http databar https://mcp.databar.ai/mcp \
  --header "Authorization: Bearer YOUR_DATABAR_API_KEY"

claude mcp add --transport http databar https://mcp.databar.ai/mcp \
  --header "Authorization: Bearer YOUR_DATABAR_API_KEY"

Start claude in the project folder and ask "How many Databar credits do I have left?" If you get a balance back, the Databar MCP is working. Optionally install Databar's prebuilt agent skills, which teach Claude the single, bulk, table and waterfall enrichment workflows:

git clone https://github.com/databar-ai/databar-mcp-server.git
cp -r databar-mcp-server/skills/* ~/.claude/skills/

git clone https://github.com/databar-ai/databar-mcp-server.git
cp -r databar-mcp-server/skills/* ~/.claude/skills/

git clone https://github.com/databar-ai/databar-mcp-server.git
cp -r databar-mcp-server/skills/* ~/.claude/skills/

git clone https://github.com/databar-ai/databar-mcp-server.git
cp -r databar-mcp-server/skills/* ~/.claude/skills/

3. Set up sending domains now, not later

Buy two or three domains close to your main one, set up SPF, DKIM and DMARC, connect the inboxes to your sending tool and start warm-up. Warm-up takes weeks, so it should run while you build everything else. The domain warm-up guide walks through the settings.

4. Run a 20-lead test

Before any real volume, test the whole chain on a small batch:

Find 20 companies that match the ICP in CLAUDE.md using Databar.
For each: get headcount and industry, find the Head of RevOps or VP Sales,
and get a verified work email with the email waterfall.
Show me the estimated credit cost first. Save results to leads-test.csv
and draft one first email per lead in drafts-test.md

Find 20 companies that match the ICP in CLAUDE.md using Databar.
For each: get headcount and industry, find the Head of RevOps or VP Sales,
and get a verified work email with the email waterfall.
Show me the estimated credit cost first. Save results to leads-test.csv
and draft one first email per lead in drafts-test.md

Find 20 companies that match the ICP in CLAUDE.md using Databar.
For each: get headcount and industry, find the Head of RevOps or VP Sales,
and get a verified work email with the email waterfall.
Show me the estimated credit cost first. Save results to leads-test.csv
and draft one first email per lead in drafts-test.md

Find 20 companies that match the ICP in CLAUDE.md using Databar.
For each: get headcount and industry, find the Head of RevOps or VP Sales,
and get a verified work email with the email waterfall.
Show me the estimated credit cost first. Save results to leads-test.csv
and draft one first email per lead in drafts-test.md

Read every row. Wrong-fit companies here mean your ICP section needs tightening, not that you need more credits.

What does the weekly routine look like?

The point of the setup is that each week costs a couple of focused sessions, not a full-time SDR job.

• Build day. Pick one segment. Ask Claude Code to pull accounts, find contacts and verified emails, and draft a short sequence. Review the list and every email yourself.

• Load day. Import the reviewed CSV into your sending tool as a paused campaign, then start it. Databar can also push rows straight to Instantly or Smartlead with an exporter if you prefer not to handle files.

• Reply time, daily. Answer replies yourself. This is the part that turns into meetings, and it is the part you should never hand to an agent.

• Review day. Paste the week's results into Claude Code and ask what the replies have in common. Write the durable lessons into the Learnings section of CLAUDE.md.

If you prefer working in a table over a terminal for any of these steps, the same Databar enrichments run in the web app. The founder's weekend guide to an enrichment stack takes that route.

How do you stop an AI agent from wasting your credits?

An agent with a data connection will happily enrich 2,000 rows because you said "find companies like this". Four guardrails keep a solo budget intact:

1. Cost before run. Put "show a cost estimate and wait for approval above N rows" in CLAUDE.md. Enrichment and waterfall prices are visible through the MCP, so Claude can do the math.

2. Small batches first. Run 10 to 20 rows, read the output, then scale.

3. Filter before enriching. Drop wrong-fit rows first. In Databar tables, run conditions skip rows that don't match, and skipped rows use no credits.

4. Watch the balance. The Databar MCP runs with spending guardrails (balance checks and per-request cost caps), and you can set low-credit email alerts in workspace settings. Results the MCP already fetched are cached for 24 hours, so rerunning the same lookup that day doesn't charge twice.

What should a solo founder not automate?

Three things stay manual. Sending: Claude drafts, you approve, the sending tool sends. Replies: a wrong automated answer to an interested prospect costs more than the time saved. Targeting changes: let Claude suggest a new segment from reply data, but decide yourself. Everything else, from list building to enrichment to first drafts, is fair game. For a wider view of what small teams use, see prospecting tools for small teams and solopreneurs.

When should you upgrade the stack or hire?

Upgrade when a real limit shows up, not before. If you run out of credits every month on lists that convert, the next Databar tier is Scale at $495/month with 50,000 credits and unlimited team members; Build allows three workspace editors. If Claude Code's usage limits interrupt your build days, move to Max. Hire your first salesperson when replies and meetings arrive faster than you can handle them. At that point your CLAUDE.md, skills and learnings become the onboarding document for the new hire.

How does this compare with other ways to start?

FAQ

Do I need to know how to code?

No. You describe what you want in plain English and Claude Code writes and runs any scripts. You do need to be comfortable opening a terminal and reading a CSV before you send anything.

Can I try this before paying for Databar?

Yes. The 14-day trial includes 100 credits, enough to connect the MCP and run the 20-lead test above on a few steps. Plan how you spend them: a verified email alone is usually 4 or more credits.

Is Claude Pro enough, or do I need Max?

Start with Pro. Claude Code is included in all paid Claude plans and shares one usage pool with your chats. Move to Max if long build sessions keep hitting the limit.

Can I use Cursor or Codex instead of Claude Code?

Yes. The Databar MCP works with any MCP client, including Cursor, Codex and Gemini CLI. The CLAUDE.md file is Claude-specific, but the same instructions work as a rules or AGENTS file in other tools.

Start with 20 leads

Write the CLAUDE.md, connect the MCP and run one small batch this week. You will learn more from 20 reviewed rows than from planning the perfect stack. Start free with the 14-day trial and 100 credits, or book a founder demo if you want help shaping your first segment.