Skip to content

Blockmachine joins the RPC cost and free-tier benches - #3152

Merged
Flotapponnier merged 1 commit into
devfrom
feat/blockmachine
Oct 22, 2026
Merged

Flotapponnier merged 1 commit into
devfrom
feat/blockmachine

Conversation

@Flotapponnier

Copy link
Copy Markdown
Collaborator

Adds Blockmachine to bench 282 (rpc-cost) and bench 283 (rpc-free-tier), with its logo and product registry entry.

What changes on the board, computed by running the harness on the old and new catalogues:

Cell (Ethereum) Before After
dapp 10M BlockPI $29.50 Blockmachine $25 (Pro)
dapp 100M Dwellir $299 Blockmachine $125 (Business)
dapp 1B Dwellir $1,999 Blockmachine $999 (Enterprise)
indexer 10M / 100M / 1B Chainstack+Dwellir $49 / Dwellir $299 / $1,999 Blockmachine $25 / $125 / $999
simple reads 100M / 1B BlockPI $133 / Dwellir $1,999 Blockmachine $125 / $999
trace, Solana, simple 10M unchanged unchanged

Blockmachine is 8th on trace (25.2 RU a request) and cannot serve 1B trace under its 250,000 RU/min limit. Its keyed Free plan (10,000 RU a month) ranks last on bench 283; the keyless endpoint (60 RU/min per IP) is a public endpoint, not a plan, and is not ranked, as for other keyless RPCs.

Sources: blockmachine.io/llms.txt plan table (polled, baseline hash), the Ethereum gateway method page (manual, its HTML embeds a build id), and the team's own statement that every method is 1 RU except trace and debug. The Free plan is confidence: uncertain: the pricing card says 300 requests/min with archive, llms.txt says 100 RU/min with archive from Standard up.

Checks: go test ./... in the harness, pnpm validate (253 specs), tsc --noEmit, bun test src/ (678 pass).

Before main: the harness image bakes the catalogue in, so the VPS container needs a rebuild after merge, then a worker rebuild for the spec. We told Blockmachine they would see the results before they go live, so staging is the review copy.

… 15 cost cells

Blockmachine bills 1 Request Unit for almost every method, archive reads
included, and sells 20M units for $25, 100M for $125 and 1B for $999. On
that it leads every Ethereum dapp and indexer cell and simple reads from
100M up. It wins no trace cell: debug and trace cost 25 to 50 units, and a
billion trace requests need 9,722 RU/s against its 4,167 ceiling.

Catalogue entry from its own llms.txt plan table and the Ethereum gateway
method page, with the archive rule confirmed by the team on 2026-10-22
("1 RU except for trace and debug"). Its pages disagree on the Free plan's
rate limit and archive access, so that plan is marked uncertain and the
conflict is written into both specs. The keyless endpoint is a public
endpoint, not a plan, and is not ranked.

Spec prose recomputed from the harness on both catalogues: winners, the
Alchemy and QuickNode ratios (now 1.7x to 16.8x), provider counts, the
free-tier spread (OnFinality 12.0M to Blockmachine 10,000 simple reads).
Logo, product registry entry and the committed cost curve included.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant