Sales outreach · checked 2026-10-19Can I vibecode Gojiberry AI?
NOT REALLY· don't bother
Gojiberry is two things: a source of buying signals (people who engaged with a competitor, followed your page or changed jobs) and a machine that contacts them through LinkedIn seats and email senders. Both halves lean on things a coding session can't produce. LinkedIn has no open API for that activity, so signals mean scraping or a data vendor, and automating LinkedIn actions can get an account restricted. Cold email at volume needs warmed domains and mailboxes, and contact data comes from a waterfall of paid providers. What you can build is the middle: score a list you already have against your ideal customer and write messages you approve before they go out.
$99/moprice
$1,188it costs you a year
weekendbuild time
—people replaced it
What you need on hand.
DifficultyAdvanced · Real infrastructure or OAuth work.
TimeA weekend to build; a new sending domain needs weeks of warm-up before volume
- Claude Code, Codex or CursorAny of them can run the prompt.
- Node.js 22better-sqlite3 compiles a native module, so build tools must be installed.
- A sending domain and mailboxUse a separate domain from your main one; set SPF, DKIM and DMARC before the first send.
- An LLM API keyAny OpenAI-compatible endpoint works, including a local model.
- A lead listYour Gojiberry contacts exported as CSV, or a list from sources you may lawfully use.
Before you buildGojiberry advertises a free trial: use it to see whether the signals fit your market before building anything.
- 1. Install one agent: Claude Code, Codex, Cursor.
- 2. Make an empty folder and open it in the agent.
- 3. Paste the prompt below, press enter, and answer its questions as it builds.
The prompt
Build me a local outbound desk to replace the part of Gojiberry AI that I can run myself. Requirements: - Node.js 22 + TypeScript: a Fastify server with Eta pages plus a CLI. `npm start` serves http://localhost:4750; `outbound score` and `outbound send --dry-run` run from the terminal. - Inputs: ./icp.md (who I sell to), my website URL (fetched once and summarised by the LLM) and leads imported from CSV (name, title, company, domain, email, LinkedIn URL, a free-text "signal" column). No scraping of LinkedIn or any other site. - Storage: better-sqlite3 in ./data/outbound.db: leads, scores, drafts, sends, replies and a suppression list. - Scoring: one call per lead to an OpenAI-compatible endpoint (LLM_BASE_URL, LLM_API_KEY, LLM_MODEL in .env) with a fixed rubric: fit 0 to 5, signal strength 0 to 5, one-sentence reason. Store the rubric version; leads under a threshold are parked. - Drafting: for leads above the threshold write a 60 to 90 word first email that names the signal, uses only facts from the CSV and the site summary, and ends with one question. Flag any draft that mentions something not in those inputs. - Approval queue: I read, edit and approve every draft. Nothing sends without approval. - Sending: nodemailer SMTP from .env, at most DAILY_CAP emails a day (default 20), 60 to 180 seconds apart, only in the recipient's working hours (`luxon`), with a List-Unsubscribe header and a footer line. Check the suppression list before every send. - Replies: poll IMAP (`imapflow`) every 10 minutes, match by Message-ID, stop follow-ups on a reply, and add bounces and "unsubscribe" replies to the suppression list. - Follow-ups: at most two, 4 and 9 days after the first email, drafted and approved the same way. - Single user. No accounts, no telemetry, local-first. Outbound calls: the LLM endpoint, SMTP and IMAP, and my website URL. - Out of scope: LinkedIn automation, signal detection, email finding and enrichment, warm-up. - README: SPF, DKIM and DMARC on a separate sending domain, a note that cold-email consent rules differ by country, and that LinkedIn's terms forbid automation.
Opens your agent with the prompt filled in, you press enter. Or copy it and paste it into an empty folder.
Then keep going
Next prompts, same folder.
The first prompt builds the core. Run these one at a time in the same folder to grow it into your own Gojiberry AI.
Add reply handling to the outbound desk in this folder. Requirements: - Label each incoming reply with the LLM as interested, not now, wrong person, referral, unsubscribe or out of office, and store the label with a one-line reason. - For interested, not-now and referral replies, draft an answer into the approval queue; never send automatically. - Out-of-office replies pause follow-ups until the date in the message when one is present. - Show a reply inbox page with filters per label. - Update the README.
Add reporting to the outbound desk in this folder. Requirements: - A weekly email and a report page: emails sent, replies, positive replies, bounces and unsubscribes, grouped by score band, job title and company size when present. - Show counts next to every rate and print "not enough data" below 30 sends in a group. - Include the rubric version used for scoring, so rubric changes can be compared. - Export the same tables as CSV. - Update the README.
Add rubric comparison to the outbound desk in this folder. Requirements: - Keep rubrics as versioned markdown files; a `outbound rescore --rubric v2` command scores the same leads again without touching sends. - A page that shows, for each lead, both scores and where they disagree by 2 or more points. - After 4 weeks, report reply rate per rubric for leads each rubric would have sent, and label small samples clearly. - Update the README.
What you lose
- Signal detection: competitor-post engagement, page followers, job changes and funding found for you
- The email and phone enrichment waterfall across 15+ providers
- LinkedIn connection requests and messages sent from a seat in the same sequence as email
- A unified inbox with drafted replies and lead scoring across both channels
- Website visitor identification
Why people still pay
Founders and small sales teams pay for prospects that arrive already filtered by intent, without building lists by hand or stitching several tools together. One subscription bundles signals, enrichment, sending and an inbox, which is otherwise a data tool, a sequencer and the time to run both.
Gojiberry AI or your own copy?
Over
Your usage
12h ×$/h
Gojiberry AI, 1 yr$1,188
Your own copy, 1 yr$116running + build session
You keep$1,073pays for itself in the first month
Estimate · Our estimate: a domain and one mailbox cost a few dollars to about ten a month, and LLM tokens for a few hundred leads are a few dollars. Enrichment and signal data are not included and would be paid separately. The build is token usage on your agent plan.
Take your data, then cancel.
1 · Take your data out of Gojiberry AI
- 1Contacts: the Contact section has an Export button at the top, according to SmartReach's guide to connecting Gojiberry; we didn't find the step in Gojiberry's own articles.
- 2The help center has a Technical & API collection covering the API, webhooks and authentication, and an article on the MCP server: use them to pull leads and campaign data programmatically.
- 3The terms describe no export tool and say terminated accounts lose access to services and stored content, so copy what you need first.
2 · Cancel Gojiberry AI
- 1The Billing & plans collection has How to Cancel Your Subscription and How to Cancel an Add-on; we couldn't read the steps, so open them in the billing tab.
- 2The terms say you can cancel at any time and the plan stays active until the end of the billing cycle.
- 3Add-ons (extra agents, lead sources, organization members) are separate subscriptions: cancel each one.
- Fees are non-refundable and there are no partial-period refunds, so cancel before the renewal date; the help center has an article for the case where you forgot and were charged.
- Disconnect your LinkedIn account from Gojiberry as part of leaving (see the Connecting LinkedIn collection), and confirm no sequences are still running.
- The terms make you responsible for following the rules of LinkedIn and other platforms, and Gojiberry isn't liable for restrictions on your accounts.
- Your email domain and warmed mailboxes are yours: keep them if you move to another sender.
Checked 2026-10-19 against Gojiberry AI's docs: 1, 2, 3, 4, 5
The escape hatch
Don't feel like building it? These are already free.
LinkiFreeSelf-hosted AI SDR with LinkedIn sequences, cold email and lead enrichment; source-available under a Sustainable Use licence, free for your own use but not OSI open source.★ 168 · active Oct 2026 · docker to self-host
WebVisit ↗LinkedGrowOpen sourceAGPL-3.0 agents that find prospects under competitor posts and searches, score fit and message them, self-hosted with Docker Compose; small project, and LinkedIn automation risks your account.★ 36 · active Oct 2026 · docker to self-host
WebVisit ↗
WebVisit ↗LinkedGrowOpen sourceAGPL-3.0 agents that find prospects under competitor posts and searches, score fit and message them, self-hosted with Docker Compose; small project, and LinkedIn automation risks your account.★ 36 · active Oct 2026 · docker to self-host
WebVisit ↗
The numbers
Gojiberry AI pricing
| Plan | Monthly | Annual, per month | What you get |
|---|---|---|---|
| Pro | $99 | — | 2 AI agents, 2 LinkedIn seats, 200 credits a month, up to 1,800 prospects contacted a month, warm-lead sourcing, website visitor identification, unified inbox, lead scoring, email and phone enrichment from 15+ providers, CRM, API and MCP integrations. |
| Custom | custom | — | A custom number of agents and prospects, more senders, a dedicated customer success manager, custom integrations and an admin and team view. |
- Billing
- Monthly subscription that renews automatically. The terms say you can cancel at any time and the plan stays active to the end of the billing cycle, and that fees are non-refundable except where the law requires, with discretionary refunds for duplicate charges and billing errors. Changes to plan and add-ons can be made from the billing settings.
- Hidden costs
- Extra capacity is sold as add-ons: an AI Outreach Agent is $59 a month each (one audience, one LinkedIn sender, about 900 contacted prospects a month), an Automated Lead Source is $29 a month each (two are included), and an Organization Member is $99 a month each. Credits cover enrichment, AI chat and faster agent runs; extra credit packs start at $25 a month.
- Source
- gojiberry.ai
Also on the list
Related apps
A free trial is advertised on the site; the pricing section and help center don't state its length or limits, and a help article covers what happens when it ends.
Not really. Founders and small sales teams pay for prospects that arrive already filtered by intent, without building lists by hand or stitching several tools together. One subscription bundles signals, enrichment, sending and an inbox, which is otherwise a data tool, a sequencer and the time to run both. The prompt on this page builds the closest honest substitute.
About $99 a month on the Pro plan, $1,188 a year (checked 2026-10-19).
Signal detection: competitor-post engagement, page followers, job changes and funding found for you; The email and phone enrichment waterfall across 15+ providers; LinkedIn connection requests and messages sent from a seat in the same sequence as email; A unified inbox with drafted replies and lead scoring across both channels; Website visitor identification. If any of those matter to you, keep paying.
Yes: Linki (free), LinkedGrow (open source).