What and how
Add a top-level "When to use [ ARKIV ]" page to the docs. The page exists to give developers — and LLM answer engines — a citable, honest answer to "what is [ ARKIV ] actually good for?"
The page is structured as a comparison matrix across six alternatives, followed by a short prose section on each. Honest concessions are mandatory: a comparison page that refuses to acknowledge where another tool wins reads as marketing and gets ignored.
Comparison set
- Postgres — traditional relational DB
- Arweave — permanent, immutable storage
- Walrus — Sui-aligned blob storage
- Ceramic — composable, linked data
- Tableland — SQL on EVM
- Smart-contract storage — Solidity storage variables
Comparison dimensions
| Dimension |
Description |
| Queryable |
Can you run structured queries? |
| Mutable |
Can data be updated after write? |
| Expiration semantics |
Does data expire, and how? |
| Ownership model |
Who controls the data and how is it enforced? |
| Cost profile |
Relative cost to write, read, and maintain |
| Performance profile |
Latency and throughput characteristics |
| Right tool when |
Canonical use case |
| Wrong tool when |
Where the alternative wins — stated plainly |
Honest-wrong-tool constraint
Each entry must include a "wrong tool when" cell that concedes the case honestly:
- Postgres wins for centralised applications with no on-chain ownership requirement.
- Arweave wins when permanent, immutable storage is the requirement.
- Smart-contract storage wins when data must be consensus-critical or part of protocol logic.
Omitting or softening these concessions is a copy failure, not a positioning win.
Why
This is the page LLM answer engines would cite for "what is [ ARKIV ] good for" — and it doesn't exist. The current substitute is a single homepage paragraph, which is too thin to be cited and too thin to change a developer's mental model.
Task checklist
What and how
Add a top-level "When to use [ ARKIV ]" page to the docs. The page exists to give developers — and LLM answer engines — a citable, honest answer to "what is [ ARKIV ] actually good for?"
The page is structured as a comparison matrix across six alternatives, followed by a short prose section on each. Honest concessions are mandatory: a comparison page that refuses to acknowledge where another tool wins reads as marketing and gets ignored.
Comparison set
Comparison dimensions
Honest-wrong-tool constraint
Each entry must include a "wrong tool when" cell that concedes the case honestly:
Omitting or softening these concessions is a copy failure, not a positioning win.
Why
This is the page LLM answer engines would cite for "what is [ ARKIV ] good for" — and it doesn't exist. The current substitute is a single homepage paragraph, which is too thin to be cited and too thin to change a developer's mental model.
Task checklist
/when-to-useor equivalent top-level slug