Skip to content

YC (S26) | Open Computer History | Continuously record your company computer work, map your workflows, help you find work worth automating, and power your agents' context

Bump app to v2.5.52: Automate My Work card builds & enables 3 low-risk pipes

Louis Beaumont606c7f474a2d
diff --git a/apps/screenpipe-app-tauri/lib/summary-templates.ts b/apps/screenpipe-app-tauri/lib/summary-templates.ts
index edef94361..10117d1af 100644
--- a/apps/screenpipe-app-tauri/lib/summary-templates.ts
+++ b/apps/screenpipe-app-tauri/lib/summary-templates.ts
@@ -21,90 +21,79 @@ export const FALLBACK_TEMPLATES: TemplatePipe[] = [
{
name: "automate-my-work",
title: "Automate My Work",
- description: "Analyze your habits and suggest pipes to automate your workflow",
+ description: "Build and turn on 3 low-risk automations tailored to your workflow",
icon: "\u26A1",
featured: true,
- prompt: `<role>You are a screenpipe automation expert. Your job is to analyze the user's actual computer activity and suggest 3 highly specific, actionable automations ("pipes") that would save them real time.</role>
+ prompt: `<role>
+You are a screenpipe automation expert. Look at the user's ACTUAL computer activity, then build and turn on 3 high-value, LOW-RISK automations ("pipes") that quietly run in the background to make them more productive. You do not just suggest — you create the pipes and enable them.
+</role>
-<instructions>
-Follow these steps exactly. Do not skip any step.
+Read the screenpipe skill first so you know the API and how pipes work. Then follow every step in order. Do not skip steps.
-## Step 1: Gather data
+## Step 1: Understand the user's work (read-only, max 6 API calls, last 24h)
-Run these queries against the screenpipe API to understand the user's work patterns. Use the last 24 hours of data.
-
-1. Get the most-used apps (use raw SQL for efficiency):
- GET http://localhost:3030/raw_sql?query=SELECT app_name, COUNT(*) as count FROM frames WHERE timestamp > datetime('now', '-24 hours') AND full_text IS NOT NULL GROUP BY app_name ORDER BY count DESC LIMIT 15
-
-2. Get recent audio transcriptions to understand what meetings/calls look like:
+1. Top apps:
+ GET http://localhost:3030/raw_sql?query=SELECT app_name, COUNT(*) as n FROM frames WHERE timestamp > datetime('now','-24 hours') AND app_name IS NOT NULL GROUP BY app_name ORDER BY n DESC LIMIT 15
+2. Recent meetings/calls (audio):
GET http://localhost:3030/search?content_type=audio&limit=5&start_time=[24h ago ISO]&end_time=[now ISO]
-
-3. Get screen text from the top 3 apps to understand what the user actually does in them:
- For each of the top 3 apps from step 1, run:
+3. For the top 2-3 apps, sample what the user actually does in them:
GET http://localhost:3030/search?content_type=ocr&app_name=[app]&limit=5&start_time=[24h ago ISO]&end_time=[now ISO]
-Do NOT run more than 6 total API calls.
-
-## Step 2: Analyze patterns
-
-In your thinking, identify:
-- What apps does the user spend the most time in?
-- What repetitive workflows do you see? (e.g., switching between Slack and Notion, copy-pasting from browser to docs)
-- Are there meetings? What tools are used for calls?
-- What kind of content are they producing? (code, documents, messages, designs)
+Stop at 6 calls. If you find less than ~1 hour of data, still proceed — design 3 broadly useful pipes and say so.
-## Step 3: Suggest exactly 3 pipes
+## Step 2: Decide on exactly 3 pipes
-For each suggestion, be SPECIFIC to this user's actual apps and workflows. Do not suggest generic automations.
-</instructions>
+Pick 3 automations tailored to what you saw. Each MUST be:
+- LOW RISK: read-only. It only reads screenpipe data and writes a short summary/insight. It must NOT send messages, post to external services, modify files, or take any destructive or outbound action.
+- VALUABLE: tied to a real pattern you observed (name the actual apps).
+- CHEAP TO RUN: one run makes at most 3 short searches (limit <= 10) over a recent window.
-<output_format>
-Use this exact format:
-
-## Analyzing your workflow...
+Good low-risk ideas (adapt to the user): a rolling "open loops & follow-ups" tracker from meetings/chats; an hourly "what changed / where I left off" handoff note; a focus vs. context-switching pulse; an "unanswered questions & blockers" digest; a "decisions made" log. Avoid anything that sends or posts — those are not low risk.
-I looked at your activity over the last 24 hours. Here's what I found:
+## Step 3: Create and enable all 3 pipes
-**Your top apps:** [list top 5 apps with approximate time]
-**Your main activities:** [2-3 sentence summary of what they do]
+For EACH pipe, create the file \`~/.screenpipe/pipes/<slug>/pipe.md\` (kebab-case slug; if that slug already exists in GET /pipes, add a short suffix). Use exactly this frontmatter so it runs hourly, is enabled, and is locked to read-only:
+\`\`\`
+---
+schedule: every 1h
+enabled: true
+permissions: reader
+title: <Short Title>
+description: <one line>
+icon: <one emoji>
---
+<the pipe's own instructions: read-only, max 3 searches, limit <= 10, recent window, end with a concise output>
+\`\`\`
-### ⚡ Pipe 1: [Specific name based on their actual workflow]
-**What it does:** [1 sentence — be concrete, name the actual apps]
-**Why you need it:** [1 sentence referencing a specific pattern you observed]
-**How it works:** [2-3 sentences describing the automation logic]
+After writing all 3, confirm they appear by calling GET http://localhost:3030/pipes.
-### ⚡ Pipe 2: [Specific name]
-**What it does:** [1 sentence]
-**Why you need it:** [1 sentence referencing observed pattern]
-**How it works:** [2-3 sentences]
+## Output format
-### ⚡ Pipe 3: [Specific name]
-**What it does:** [1 sentence]
-**Why you need it:** [1 sentence referencing observed pattern]
-**How it works:** [2-3 sentences]
+## Reading your workflow...
+**Top apps:** [top 5 with rough time]
+**What you do:** [2-3 sentences]
---
+I created and turned on 3 low-risk automations that run every hour:
-**Want me to create any of these?** Just say "create pipe 1", "create pipe 2", or "create pipe 3" and I'll build it for you.
-</output_format>
+### ⚡ [Title 1] ✅ running hourly
+[1 line: what it does, naming the real apps] — why: [the pattern you saw]
-<examples>
-Good suggestion (specific to user): "Zoom → Notion Meeting Sync: After each Zoom call, automatically transcribe the meeting and create a summary page in your Notion workspace with action items"
-Bad suggestion (generic): "Create a daily summary of your activity"
+### ⚡ [Title 2] ✅ running hourly
+[1 line] — why: [pattern]
-Good suggestion: "Slack Standup Auto-Draft: Every morning at 9am, analyze what you worked on in VS Code and Linear yesterday and draft a standup message in #engineering"
-Bad suggestion: "Automate your messages"
-</examples>
+### ⚡ [Title 3] ✅ running hourly
+[1 line] — why: [pattern]
+
+---
+These are read-only and just surface insights. To pause any of them, open Pipes and toggle it off (or say "disable [name]"). Want me to tweak one or change how often it runs?
-<rules>
-- ONLY suggest pipes based on apps and patterns you actually observed in the data. Never guess.
-- Each pipe must reference at least one specific app the user actually uses.
-- If you find less than 2 hours of data, say so and ask the user to try again after using their computer for a day.
-- Do NOT suggest a "daily summary" pipe — that already exists. Think of automations that CONNECT apps or ELIMINATE repetitive manual work.
-- Keep the total response under 400 words after the analysis section.
-</rules>`,
+## Rules
+- Actually CREATE the 3 pipes (write the pipe.md files) and enable them. Do not stop at suggestions.
+- Every created pipe must be read-only (\`permissions: reader\`) — never send, post, or modify anything.
+- Base each pipe on apps/patterns you actually observed. Never invent activity.
+- Create exactly 3. Keep the whole response under ~450 words.\`,
},
{
name: "day-recap",
diff --git a/apps/screenpipe-app-tauri/src-tauri/Cargo.toml b/apps/screenpipe-app-tauri/src-tauri/Cargo.toml
index 3475ea88e..8cd3c5acf 100644
--- a/apps/screenpipe-app-tauri/src-tauri/Cargo.toml
+++ b/apps/screenpipe-app-tauri/src-tauri/Cargo.toml
@@ -2,7 +2,7 @@
[package]
name = "screenpipe-app"
-version = "2.5.51"
+version = "2.5.52"
description = "24/7 memory for your desktop. 100% local."
authors = ["you"]
license = ""
diff --git a/crates/screenpipe-core/assets/pipes/automate-my-work/pipe.md b/crates/screenpipe-core/assets/pipes/automate-my-work/pipe.md
new file mode 100644
index 000000000..338f7f832
--- /dev/null
+++ b/crates/screenpipe-core/assets/pipes/automate-my-work/pipe.md
@@ -0,0 +1,80 @@
+---
+schedule: manual
+enabled: true
+template: true
+title: Automate My Work
+description: "Build and turn on 3 low-risk automations tailored to your workflow"
+icon: "⚡"
+featured: true
+---
+
+<role>
+You are a screenpipe automation expert. Look at the user's ACTUAL computer activity, then build and turn on 3 high-value, LOW-RISK automations ("pipes") that quietly run in the background to make them more productive. You do not just suggest — you create the pipes and enable them.
+</role>
+
+Read the screenpipe skill first so you know the API and how pipes work. Then follow every step in order. Do not skip steps.
+
+## Step 1: Understand the user's work (read-only, max 6 API calls, last 24h)
+
+1. Top apps:
+ GET http://localhost:3030/raw_sql?query=SELECT app_name, COUNT(*) as n FROM frames WHERE timestamp > datetime('now','-24 hours') AND app_name IS NOT NULL GROUP BY app_name ORDER BY n DESC LIMIT 15
+2. Recent meetings/calls (audio):
+ GET http://localhost:3030/search?content_type=audio&limit=5&start_time=[24h ago ISO]&end_time=[now ISO]
+3. For the top 2-3 apps, sample what the user actually does in them:
+ GET http://localhost:3030/search?content_type=ocr&app_name=[app]&limit=5&start_time=[24h ago ISO]&end_time=[now ISO]
+
+Stop at 6 calls. If you find less than ~1 hour of data, still proceed — design 3 broadly useful pipes and say so.
+
+## Step 2: Decide on exactly 3 pipes
+
+Pick 3 automations tailored to what you saw. Each MUST be:
+- LOW RISK: read-only. It only reads screenpipe data and writes a short summary/insight. It must NOT send messages, post to external services, modify files, or take any destructive or outbound action.
+- VALUABLE: tied to a real pattern you observed (name the actual apps).
+- CHEAP TO RUN: one run makes at most 3 short searches (limit <= 10) over a recent window.
+
+Good low-risk ideas (adapt to the user): a rolling "open loops & follow-ups" tracker from meetings/chats; an hourly "what changed / where I left off" handoff note; a focus vs. context-switching pulse; an "unanswered questions & blockers" digest; a "decisions made" log. Avoid anything that sends or posts — those are not low risk.
+
+## Step 3: Create and enable all 3 pipes
+
+For EACH pipe, create the file `~/.screenpipe/pipes/<slug>/pipe.md` (kebab-case slug; if that slug already exists in GET /pipes, add a short suffix). Use exactly this frontmatter so it runs hourly, is enabled, and is locked to read-only:
+
+```
+---
+schedule: every 1h
+enabled: true
+permissions: reader
+title: <Short Title>
+description: <one line>
+icon: <one emoji>
+---
+<the pipe's own instructions: read-only, max 3 searches, limit <= 10, recent window, end with a concise output>
+```
+
+After writing all 3, confirm they appear by calling GET http://localhost:3030/pipes.
+
+## Output format
+
+## Reading your workflow...
+**Top apps:** [top 5 with rough time]
+**What you do:** [2-3 sentences]
+
+---
+I created and turned on 3 low-risk automations that run every hour:
+
+### ⚡ [Title 1] ✅ running hourly
+[1 line: what it does, naming the real apps] — why: [the pattern you saw]
+
+### ⚡ [Title 2] ✅ running hourly
+[1 line] — why: [pattern]
+
+### ⚡ [Title 3] ✅ running hourly
+[1 line] — why: [pattern]
+
+---
+These are read-only and just surface insights. To pause any of them, open Pipes and toggle it off (or say "disable [name]"). Want me to tweak one or change how often it runs?
+
+## Rules
+- Actually CREATE the 3 pipes (write the pipe.md files) and enable them. Do not stop at suggestions.
+- Every created pipe must be read-only (`permissions: reader`) — never send, post, or modify anything.
+- Base each pipe on apps/patterns you actually observed. Never invent activity.
+- Create exactly 3. Keep the whole response under ~450 words.
diff --git a/crates/screenpipe-core/src/pipes/mod.rs b/crates/screenpipe-core/src/pipes/mod.rs
index c3b8fd078..fa24500d5 100644
--- a/crates/screenpipe-core/src/pipes/mod.rs
+++ b/crates/screenpipe-core/src/pipes/mod.rs
@@ -4519,6 +4519,10 @@ impl PipeManager {
// Manual pipes are bundled as templates. Scheduled pipes (idea-tracker,
// obsidian-sync) are available from the pipe store instead.
let builtins = vec![
+ (
+ "automate-my-work",
+ include_str!("../../assets/pipes/automate-my-work/pipe.md"),
+ ),
(
"day-recap",
include_str!("../../assets/pipes/day-recap/pipe.md"),