v0.1 · macOS (Linux + Windows soon)

Build pipelines like it's your computer.
Because it is.

Pipes.Run is a desktop app for engineers who don't want a SaaS sitting between them and their data. Build, debug, and run pipelines locally. Bring your own Ollama for AI nodes. Push to our cloud only when production calls for it.

Pay once, own it No telemetry Works offline
Save as Pipeline
AI funding research · Pipes.Run
D

Find AI funding announcements from this week. Pull company, amount, and round stage.

P

Searched 12 sources, extracted 8 results:

CompanyAmountRound
Decagon$131MSeries C
Sierra AI$210MSeries C
Harvey$300MSeries E

techcrunch.com · theinformation.com · 10 more

Ask a follow-up...
Grounded · 12 sources · 2.3s · local gemma3:4b via Ollama
↳ Scheduled Runs Mon 9:00 AM
→ postgres://leads
$ pipesrun --version → v0.1 Finally. A pipeline tool I don't have to log in to. Early access · 30-day refund · no card on file for desktop
Why we built this

Three things every pipeline tool got backwards.

Every data pipeline tool on the market made the same trade. They built for the cloud first, the developer second, and your laptop never. We thought that was strange — so we did the opposite.

01

Your data is sensitive.

The point of a pipeline is to transform data you care about — customer records, financials, internal documents. Why does every tool want a copy of it on someone else's servers first?

02

Your laptop is powerful.

An M-series Mac can run a 7B model and process 100k rows in seconds. Yet you're paying a SaaS $200/mo to do it remotely, slower, with rate limits and cold starts.

03

The cloud is for production.

Scheduled jobs, webhook triggers, team sharing — those genuinely need the cloud. The 90% of work that's building and iterating doesn't. Pay for what you actually need.

What's in the box

A complete pipeline IDE, installed locally.

Everything you need to build, test, and run real pipelines — without sending anything to a cloud you don't control.

Visual canvas

Drag, connect, configure. Every node shows its schema, sample data, and run time inline.

Drop into Python

Any node can be a Python script. Full sandbox, pip install, local stdlib. No DSL to learn.

Instant runs

⌘R runs the pipeline. Edit a node, hit run again. No deploys, no waits, no log tailing.

Credentials stay local

API keys live in your OS keychain. We don't have a server that could be breached.

40+ built-in nodes

Postgres, HTTP, S3, SFTP, Slack, Notion, Airtable, Stripe, Snowflake, Google Sheets, more.

Full observability

Every run logged with row counts, durations, errors, full payloads. SQLite-backed, all yours.

Version control

Pipelines are plain YAML files. Commit them. Diff them. Code-review them. Yours forever.

Native, not Electron

Built in Tauri. Fast launch, small binary, plays nicely with your battery.

One-click cloud push

When you're ready, "Deploy to Cloud" sends the YAML to our runners. Same exact pipeline.

The wedge

The only pipeline tool with local LLMs built in.

Pipes.Run ships with first-class Ollama support. Drop a Local AI node into your pipeline, point at a model on your machine, and process documents, classify text, extract structured data — without an API key, a rate limit, or your data ever touching the open internet.

  • Ollama auto-detection.If you have Ollama installed, your local models show up in the node picker.
  • Structured output, schema-first.Define the JSON shape you want. We constrain the model to return it. No regex parsing.
  • Hybrid: local + remote.Use a local model for PII redaction, then a frontier model for the heavy reasoning. We handle the handoff.
  • Zero token costs for local nodes.Run a million-row enrichment overnight. Pay your electric bill, not OpenAI.
enrich-leads.pipe.yaml
# Pipeline: read leads, enrich with local Gemma,
# write to Postgres. No data leaves the laptop.

name: enrich-leads
runtime: local   # or "cloud"

nodes:
  - id: source
    type: postgres.query
    sql: "select * from leads where enriched is null"

  - id: enrich
    type: llm.local      # 🧠 runs on your machine
    provider: ollama
    model: gemma3:4b
    prompt: "Classify industry and seniority from: {{row.bio}}"
    schema:
      industry: string
      seniority: enum[ic, manager, exec]

  - id: sink
    type: postgres.upsert
    table: leads
    key: id

edges:
  - source → enrich → sink
The flow

Three steps. No surprises.

The Pipes.Run promise: nothing leaves your laptop until you explicitly tell it to.

1

Build it locally

Drag, drop, connect, configure. Hit ⌘R. Iterate at the speed of thought.

# everything runs here
$ pipesrun run
✓ 1,204 rows · 1.8s
2

Commit it

Save the pipeline YAML to your repo. Diff it, review it, version it like real code.

$ git add enrich.pipe.yaml
$ git commit -m "add lead enrichment"
3

Deploy (optional)

When you need scheduled runs or webhook triggers, hit Deploy. Our cloud picks up the same YAML.

$ pipesrun deploy --cron "0 */6 * * *"
→ deployed · runs every 6h
Pricing

Buy the tool. Rent the runway.

The desktop app is yours forever, with a one-time fee. The cloud is optional, monthly, and only for what genuinely needs to live on a server.

Start here

Pipes.Run Desktop

Yours to keep.
$99one-time
Full app, every node, every integration. No subscription, no telemetry, no expiration.
  • Lifetime license · macOS (Apple Silicon + Intel)
  • 40+ built-in nodes & custom Python
  • Local Ollama / LM Studio support
  • One year of updates included
  • Use on up to 3 of your machines
30-day refund · no questions · continued updates after year 1: $49/yr (optional)
vs everyone else

The honest comparison.

We're not trying to be all things. We're trying to be the right thing for engineers who'd rather own their tools than rent them.

 Pipes.Runn8n CloudZapierCustom Python
Runs on your laptop Yes No No Yes
Local LLMs (Ollama) Native~ DIY No Yes
Visual pipeline builder Yes Yes Yes No
Pay once, own it Yes Subscription Subscription Free
Cloud deploy (when needed) Optional~ Required~ Required DIY
Data stays on your machine Default No No Yes
Time to first pipeline~ 5 min~ 20 min~ 10 min~ 2 hours
FAQ

Questions we get all the time.

If we missed yours, we're at hello@pipes.run and we actually reply.

Why one-time pricing instead of a subscription?
Because we're building a tool, not a service. You wouldn't pay Adobe a monthly fee to keep using Photoshop you already installed (well — you do, and we think that's a shame). The desktop app does its job whether or not we exist next year. That's a feature.
Can I schedule pipelines on my laptop?
You can — Pipes.Run includes a local scheduler that runs while the app is open. But realistically, "always-on" production schedules belong on a server. That's what Cloud is for. We're honest about this rather than pretending your MacBook is production infrastructure.
What happens if I don't renew updates after year 1?
Nothing. The app you have keeps working forever — same nodes, same features, same everything. If you want new nodes, bug fixes, and new integrations released after your update window, the renewal is $49/yr. Skip it for a year, come back later — totally fine.
Do I need Cloud to use Pipes.Run?
No. The desktop app is fully functional on its own. Cloud is purely for when you need scheduled runs, webhook triggers, or a place pipelines live while your laptop is closed. Plenty of customers buy Desktop and never touch Cloud.
Which LLM providers are supported?
Locally: Ollama, LM Studio, llama.cpp, anything OpenAI-API-compatible. Remote: OpenAI, Anthropic, Google, Mistral, Groq, OpenRouter. Bring your own keys — we never proxy through our servers.
Is it open source?
Not currently. The pipeline format (YAML) is open and documented — your pipelines work forever regardless. We may open-source the runtime in the future. The desktop app itself is closed-source so we can sustain development.
How is this different from n8n?
n8n is excellent for cloud-first SaaS automation with hundreds of integrations. Pipes.Run is for engineers who want local-first execution, native AI nodes that use their own machine, and one-time ownership of the tool. Different buyers, different jobs.
Can I get a refund?
Yes. 30 days, no questions, no friction. Email hello@pipes.run with your order number and we'll process it within 24 hours. We'd rather you not use the app than feel stuck with it.

Pay once.
Own it forever.

Pipes.Run is the data pipeline tool we wished existed: built local-first, owned outright, cloud-deployable when it matters. $99, lifetime, 30-day refund.

Buy Pipes.Run · $99 once