Solana grants list: how to map the ecosystem's funding programs from public data
What people are actually looking for
Search for a Solana grants list and the need behind it is usually one of three: a builder deciding where to apply, a researcher tracking which sponsors fund what, or an analyst watching how work markets in the ecosystem shift. No single page answers any of them, because the ecosystem funds work through separate channels: per-program grant tracks (application forms, milestone reviews), hackathon prize pools, and time-boxed bounties posted on public boards. The largest public board for the Solana ecosystem is Superteam Earn, where sponsors post bounties and projects with deadlines, token labels, and submission counts, all behind a readable public API.
The durable move is to stop hunting for someone else's static list and build the mapping from rows. A snapshot gives you the fields (sponsor, type, deadline, status, automation policy) and the history that a live board overwrites every week.
The data source: the Superteam Earn API, flattened
The Superteam Earn dataset documented here was harvested from the public API on 2026-09-04 and ships three research files plus harvest notes:
- earn-listings-cards.json, 51 card records: the raw paginated cards dump. The same live listings repeat once per API page, so 17 unique listings appear 3 times each. Count on the
idcolumn, not raw rows. - earn-listings-full-descriptions.json, 28 listings with full description HTML. The cards API truncates this field; the descriptions carry scope, deliverables, and judging criteria. 16 of the 17 live listings appear here, plus 12 that had rotated off the live feed before harvest day.
- earn-api-endpoints.txt, 186+ reverse-engineered routes across 34 path groups: listings, comments, submissions, grants and tranches, hackathon feeds, homepage stats, sponsor operations. This is where the grants-shaped routes live, separate from the bounty routes.
- earn-README.md, the harvest notes: date, source, file inventory.
Column reference for the cards file, with the fields a mapping needs: title, type (bounty or project), token, deadline (ISO 8601), status, agentAccess (HUMAN_ONLY or AGENT_ALLOWED), sponsor, and _count (comments and submissions). The full table is in the Superteam Earn data guide.
What the 2026-09-04 snapshot shows
Seventeen unique listings were open on harvest day. The splits that matter for mapping, at row level and after dedupe:
| Split | 51 rows (raw) | 17 listings (deduped) |
|---|---|---|
| agentAccess: HUMAN_ONLY | 45 | 15 |
| agentAccess: AGENT_ALLOWED | 6 | 2 |
| type: bounty | 45 | 15 |
| type: project | 6 | 2 |
| token: USDC | 39 | 13 |
| token: USDG | 12 | 4 |
| status: OPEN | 51 | 17 |
Row-level and unique counts agree here because the two project-type listings are also the two AGENT_ALLOWED listings. The 6 AGENT_ALLOWED rows are 2 unique listings: the ZNS Solana Creator Challenge (ZNS Connect) and the Steve Agent Arena (OOBE Protocol). Both are automation-themed work, which is what the flag promises: listings where tooling-assisted submission is welcome.
The mapping method, step by step
- Dedupe on id or slug first. Every analysis after this depends on one-row-per-listing. The 51/17 gap is page structure, not market depth.
- Group by sponsor. The sponsor field turns the listing pile into an ecosystem map: which teams and communities post work, how many listings each runs, and whether they favor bounties or projects. Track the same groupings across snapshots to see sponsor activity over time.
- Keep type separated. Bounties (per-deliverable, deadline-bound) and projects (longer engagements) are different funding shapes. Mixing them flattens the comparison the way mixing grants with bounties would.
- Use deadlines for cadence. ISO 8601 timestamps sort cleanly, so a week-over-week snapshot shows how long listings stay open and how deadline density moves around hackathon seasons.
- Read submissions as crowding.
_count.submissionnext to each listing shows how contested each slot was on harvest day. As research, it is a baseline for comparing engagement tiers; it says nothing about which submission a sponsor will pick. - Flag the automation layer. agentAccess is the field other Solana boards rarely publish. A mapping that tracks AGENT_ALLOWED listings over time is tracking where agent-friendly work concentrates in the ecosystem.
- Go to the route map for the grants-shaped slice. The 186-route endpoint map gives grants, tranches, and hackathons their own groups. Anyone extending this mapping beyond the bounty board starts there, against each program's own endpoints.
Querying the snapshot without downloading it
The dataset-mcp server reads the full released files and filters them in memory; setup is one npx line on the For agents page. The layer question (MCP vs raw endpoints) is covered in the MCP vs API guide. One real call over the cards file:
get_stats("earn-bounties", { file: "earn-listings-cards.json", column: "agentAccess" })
Verbatim output from the September 2026 run of server v1.1.0:
{
"slug": "earn-bounties",
"file": "earn-listings-cards.json",
"rows": 51,
"column": "agentAccess",
"non_empty": 51,
"unique_count": 2,
"top_values": [
{ "value": "HUMAN_ONLY", "count": 45 },
{ "value": "AGENT_ALLOWED", "count": 6 }
]
}
query_dataset with a where clause on agentAccess returns the 6 AGENT_ALLOWED rows themselves; dedupe on id from there for the 2 unique listings.
What this data cannot tell you
A snapshot is one day. It records 17 open listings on 2026-09-04 and 12 descriptions that had already aged out; it does not capture listings posted after harvest, and the cards feed itself is a rolling window. Grants proper, the milestone-funded programs run by foundations and protocol teams, mostly live outside this board entirely: the endpoint map documents where those routes sit, but this dataset holds the bounty-board slice. Rewards are recorded as amounts and token labels for completeness; comparing them across sponsors or extrapolating them into earnings estimates is bad method, and the submission counts prove nothing about acceptance odds. This page maps funding programs for research. It is not investment advice, not a yield estimate, and not a recommendation to buy or sell anything.
Get the snapshot
Start with the free 10-card sample to check the schema, then take the full pack for all 51 card records, 28 full descriptions, the 186-route endpoint map, and the harvest README. Related reading: the full Superteam Earn data guide, the Solana ecosystem data overview, the web3 bounty data overview, and the tracker comparison in bounty tracker alternative.
Checkout and download run through Getly. Source: public Superteam Earn API, harvested 2026-09-04. Personal and commercial use; no resale of the files as-is.
Solana grants list questions
Where can I find a list of Solana grants?
No single official list exists. Programs run per sponsor, and the public concentration point for bounties and projects is Superteam Earn. The research route is snapshots: this dataset gives the board's slice as rows, and the endpoint map's grants and hackathon route groups show where program-level funding lives. Combine both with each program's own pages for a full picture. For the other data layers around this board (chain state, validator stats, venue intelligence), see the Solana ecosystem data overview.
What does AGENT_ALLOWED mean?
The listing's automation policy: submissions through tooling and agents are welcome, versus HUMAN_ONLY where a person must submit. In the 2026-09-04 snapshot: 6 of 51 card rows, which dedupe to 2 unique listings, both automation-themed work.
Is a grants list the same as a bounty list?
No. Grants are program funding with application routes and usually milestones; bounties are per-deliverable payouts with deadlines and submission counters. The type column (bounty vs project) and the separate grants route groups in the endpoint map keep the two apart, and any serious mapping should keep them separated.