# Product teardown

A teardown of one product from its public pages, docs and reviews.

Requires web search, through Exa. Requires web fetch.

## Install

```sh
npx @parcha/agentrun template install github:Parcha-ai/agentrun-templates/templates/product-teardown@242f29da8317fbe51d0cce4628bdb5ab61d7ccc9 --trust
```

Or open `<your AgentRun address>/#/templates/install?src=github%3AParcha-ai%2Fagentrun-templates%2Ftemplates%2Fproduct-teardown%40242f29da8317fbe51d0cce4628bdb5ab61d7ccc9`. Pinned to commit 242f29da8317fbe51d0cce4628bdb5ab61d7ccc9.

## What it does

A teardown of one product from its public pages, docs and reviews, for product people studying a competitor or a company they admire. It covers who it is for, onboarding, the core workflow, pricing and packaging with the date each price was seen, growth loops (observed or inferred), integrations, release cadence, what it does unusually well and what users complain about.

- Skill `product-teardown`: Tear down one product from public material. Fix which product it is, read who the copy addresses, reconstruct onboarding and the core workflow from docs and walkthroughs, record pricing and packaging with the date seen, find growth loops (observed or inferred), integrations and release cadence, and mine reviews for what it does well and what users complain about. Use for "teardown of", "how does X onboard", "study this product".

## What starts it

- By hand, any time: from the console, the API or the CLI.

## Inputs

| Name | Type | Required | What it is |
| --- | --- | --- | --- |
| `product` | string, 2 to 160 characters | yes | The product to tear down. Example: `Linear`. |
| `website` | string, uri | no | Its official site. This decides which product is meant when names collide. Example: `https://linear.app`. |
| `angle` | string, 0 to 300 characters | no | A part to go deep on, such as onboarding for teams or pricing against seat-based rivals. Other sections stay brief. Example: `How they convert free teams to paid`. |

## What it needs

- Harness: pi, claude-agent, codex
- Model: balanced tier, medium effort
- Tools: read, web_search, web_fetch
- Computer: Wherever your AgentRun puts a new agent
- Keys: A model key
- Learning: off

## Sample output

Input:

```json
{
  "content": "Today is 2026-11-14. Tear down this product.\n\nThe sites below are not reachable from this run. The pages below are what their URLs return today; use them in place of\nfetching, and search normally for anything else.\n\n<captured-page url=\"https://slatebird.example/\" captured=\"2026-11-14T11:00:00Z\" untrusted>\n<h1>Whiteboards your whole team actually uses</h1>\n<p>Slatebird is an online whiteboard for product and design teams. Share any board with a link; guests can comment\nwithout an account. Free for up to 3 boards; Team plan $9 per editor per month.</p>\n</captured-page>\n\n<captured-page url=\"https://slatebird.example/refer\" captured=\"2026-11-14T11:01:00Z\" untrusted>\n<h1>Refer a team</h1>\n<p>Give a friend's team one free month of Team. When they upgrade, you get one free month too. Up to 12 months.</p>\n</captured-page>\n\n<captured-page url=\"https://slatebird.example/templates\" captured=\"2026-11-14T11:02:00Z\" untrusted>\n<h1>Template gallery</h1>\n<p>1,400 community templates for retros, user story maps and roadmaps. Every template page shows \"Made with Slatebird\"\nand a \"Use this template\" button that opens a free sign-up.</p>\n</captured-page>\n\n<captured-page url=\"https://reviewhub.example/products/slatebird/reviews\" captured=\"2026-11-14T11:03:00Z\" untrusted>\n<h1>Slatebird reviews (4)</h1>\n<p>2026-10-02, 5 stars: \"Fastest whiteboard we've tried, boards load instantly.\"</p>\n<p>2026-09-18, 5 stars: \"Snappy even with 40 people on one board.\"</p>\n<p>2026-08-30, 4 stars: \"Speed is the reason we switched from our old tool.\"</p>\n<p>2026-08-11, 2 stars: \"The mobile app is clunky and crashes on my tablet.\"</p>\n</captured-page>",
  "product": "Slatebird",
  "website": "https://slatebird.example",
  "angle": "How they bring in new users"
}
```

Output:

#### Bottom line
Slatebird's clearest acquisition mechanisms are branded templates leading to signup and referrals rewarded when another team upgrades. Shared boards offer another plausible distribution path, but guest-to-user conversion is inferred, not observed. Three reviews praise speed; none establishes acquisition effectiveness. [S1][S2][S3][S4]

#### Snapshot
| Item | Finding |
|---|---|
| Company | Legal identity not publicly available |
| Category | Online team whiteboard [S1] |
| Website | https://slatebird.example/ [S1] |
| Target user | Product and design teams [S1] |
| Model | Free entry; paid editor subscriptions [S1] |
| Evidence date | All supplied pages observed 2026-11-14 |

#### Target user and positioning
“Whiteboards your whole team actually uses” emphasizes participation. The homepage names product and design teams; retrospectives, user story maps and roadmaps make the gallery's use cases concrete. Company size is not publicly available. [S1][S3]

#### Onboarding
Two entry paths are documented. A template visitor clicks “Use this template” and reaches free signup. A board guest can comment without an account. Signup details, the first authenticated screen and whether the chosen template survives signup are not publicly available. No activation time can be substantiated. [S1][S3]

#### Core workflow
The visible outline is choose a template, work on a board, share its link and receive guest comments. Working on the board is an inferred middle step: editing controls and template application are undocumented in the available material. Link sharing and account-free commenting are explicit. [S1][S3]

#### Pricing
Observed **2026-11-14**; prices come from the homepage, not a detailed pricing page. [S1]

| Tier | Published price | Known limit |
|---|---|---|
| Free | Free | Up to 3 boards |
| Team | $9 per editor per month | Other gates not publicly available |

The free cap suggests a potential expansion trigger, but the actual upgrade experience is unknown. Annual terms and price history are not publicly available. [S1]

#### Growth loops
- **Observed—referrals:** give another team one free Team month; its upgrade earns the referrer one free month, up to 12 months. The reward targets a commercial event rather than mere registration. [S2]
- **Observed—template acquisition:** the gallery advertises 1,400 community templates. Every template page carries “Made with Slatebird” and a signup button. A complete creator-contribution loop is not documented. [S3]
- **Inferred—collaboration exposure:** shared links likely introduce guests to Slatebird without registration friction. Subsequent signup prompts and conversion are unknown. [S1]
- **Inferred—free-to-paid expansion:** free boards likely encourage initial use, with the cap creating upgrade opportunities. Conversion evidence is unavailable. [S1]

#### Integrations
Integrations, API access and marketplace listings are not publicly available in the reviewed evidence. Additional searches found no matching source.

#### Releases
| Period checked | Finding |
|---|---|
| Recent public release history | No matching changelog located; cadence not publicly available |

#### What it does unusually well
Speed receives three separate positive mentions between 2026-08-30 and 2026-10-02. This is repeated praise within four supplied reviews, not a benchmark. Account-free commenting and task-specific template entry points also reduce identifiable participation and discovery friction. [S1][S3][S4]

#### Complaints
One review, dated 2026-08-11, reports clunky mobile use and tablet crashes. Treat this as an investigation lead, not a recurring complaint pattern. [S4]

#### Takeaways
Separate creator signup from collaborator participation. Tie referral rewards to meaningful outcomes. Connect templates to specific jobs, then measure the signup-to-useful-board handoff rather than celebrating registrations alone. These are analytical lessons from the documented mechanisms, not proven Slatebird results. [S1][S2][S3]

#### Assumptions and gaps
The supplied captures are the evidence base; additional searches returned unrelated products. Independent verification, acquisition metrics, detailed onboarding, product documentation, complete pricing terms, integrations and release history are not publicly available in the reviewed material. The review sample is small and its representativeness unknown.

## Tested

Passes 3 of 3 eval cases on pi, 2 of 3 on Claude Code and 3 of 3 on Codex. Not yet tried on the live web.

3 eval cases: injection-in-page, loops-and-reviews, not-public.

## Source

- Package: https://github.com/Parcha-ai/agentrun-templates/tree/242f29da8317fbe51d0cce4628bdb5ab61d7ccc9/templates/product-teardown
- Publisher: Parcha-ai
- Author: Parcha Labs
- Version: 0.1.0
- License: Apache-2.0
- Category: Research
- Page: https://marketplace.agentrun.ai/templates/product-teardown/
