CRM and sales decision
SalesTouch
Self-hosting is viable because the project's source is available (MIT); a competent engineer can stand up the MCP and run a safe, limited LinkedIn automation for a lower monthly cost than the hosted $49 plan, but you'll lose SalesTouch's managed safety infrastructure and production polish.
Visit website↗Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need — the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship.
What a replacement has to do
- Provide an MCP that: authenticate and manage LinkedIn accounts; extract prospect lists from searches/posts/groups; read profiles/posts and run LLM-based qualification; queue, rate-limit and send connection requests/messages and schedule follow-ups; record activity logs and expose an API for an external agent to call.
What it still won’t have
- SalesTouch-managed residential IP & location-matching protections
- Built-in safety defaults, automatic cooldowns and per-account heuristics
- Managed uptime, human support and ready-made analytics dashboards
- Prebuilt MCP integrations to Claude/Codex and tested agent connectors
What remains hard
- Product polish and ongoing maintenance
First-year cost
Keep paying
Paying is—cheaper in year one.
On cash alone, building overtakes the subscription at 3 seats.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a self-hosted SalesTouch MCP using Docker, Postgres, Redis, and a Linux cloud VM (DigitalOcean/AWS t3.small). Clone https://github.com/antoineDsh/SalesTouch, run the project's Docker compose, provision Postgres and Redis, configure an HTTPS reverse proxy (nginx or Caddy), and wire LinkedIn account authentication (securely store encrypted credentials). Implement LLM integration via an environment-configured OpenAI/Anthropic client; include endpoints for: extract prospects (search, post reactions, comments, group members), read profile/post, generate qualification and outreach drafts, enqueue/send LinkedIn actions, and query activity logs. Scope in: queueing, per-account daily limits, working-hours windows, cooldowns, scheduled follow-ups, basic web UI or REST endpoints for approving sensitive actions, and 90-day activity logs. Out of scope: building a custom LLM, multi-tenant billing, advanced analytics UI, or proprietary residential IP proxying. Include error handling, retries with backoff, comprehensive logging, and automated tests for the main endpoints and queue workers.
How we checked
How the score was reached
- Self-host verdict base92
- An open-source build was found+5
- 4 cited sources+3
- Price verified on pricing page+3
- 1/1 assessment runs agreed+4
- Evidence score99
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time — so the same evidence always produces the same number.
How scoring works →Cited sources · 4
Every page the run actually retrieved.
- official productSalesTouch - LinkedIn automation for AI agents
- official pricingSalesTouch Pricing - 7 Days Free
- official docsSalesTouch Features - LinkedIn Tools for AI Agents
- open sourceGodsScion/Auto_job_applier_linkedIn
Integrity checks
What held up, and what did not.