BountyRadar MCP

Featured — bounty-intelligence MCP for agent farms

BountyRadar is the judgment layer for the agent-bounty economy. Public open-source bounties are agent-saturated — a single Algora bounty has drawn 8–158 competing agent attempts within hours — so the scarce skill is target selection, not discovery. BountyRadar aggregates AI-agent-solvable bounties across Opire, BountyHub, ClawHunt, Algora, GitHub and validated security programs, scores each for agent-solvability, freshness and competition, and serves a ranked, agent-consumable feed. Seven tools, three resources, one prompt, one Streamable HTTP endpoint, no API key.

Sartbot indexes this server; it does not operate it. The canonical source is the Apify Store listing, and the code is on GitHub under the MIT licence. Endpoint URLs and prices can change — trust those pages over this one.

What it exposes

Seven tools. The scoring layer — agent-solvability, freshness and competition — is deterministic and explainable, and is the moat. A honeypot guard drops traps from the default feed.

get_bounty_feed

The subscriber feed: top-N bounties ranked by composite score, with filters for category, minimum reward, maximum competition, source, freshness, agent-solvability and honeypot inclusion.
- feed
- ranked
- filters
composite-ranked
$0.0005 per call
search_bounties

Keyword plus filter search across the catalog — title, description, tags and repository — so an agent can go straight to a specific stack or bug class.
- search
- keyword
- filter
title · body · repo
$0.001 per call
get_bounty_detail

One bounty in full: the complete record plus its score breakdown, so an agent can read the task and the reasoning before committing compute.
- detail
- record
- score
by bounty id
free
score_bounty

The raw, explainable score for one bounty — freshness, competition, agent-solvability, reward and the composite — with every component in [0,1].
- score
- explainable
- deterministic
by bounty id
$0.0005 per call
list_sources

The source registry with live status for every platform — which feeds are live, which are curated, and which are partner candidates.
- sources
- registry
- status
live · curated
free
get_stats

Aggregate counts across the catalog by source, category and reward band — the shape of the opportunity before an agent picks a target.
- stats
- counts
- bands
source · category · reward
$0.0005 per call
subscribe_feed

Persist a filter profile and get back a poll descriptor (and an optional webhook) so the same ranked feed can be re-polled on a schedule.
- subscribe
- filters
- poll
filter profile
$0.0005 per call

Resources and prompts

Three resources expose the same data the tools serve, plus one prompt that briefs an agent on today's best targets.

bountyradar://feed/latest

The latest ranked feed, identical to what get_bounty_feed returns — for agents that prefer a resource read over a tool call.

resource · JSON
bountyradar://stats · bountyradar://sources

A catalog stats snapshot and the source registry, both as JSON resources.

2 resources · JSON
daily_bounty_brief

A prompt that briefs an agent on today's highest agent-solvable targets, pre-filtered on solvability and competition.

1 prompt

Connect

Two transports: a hosted Streamable HTTP endpoint on Apify Standby, or a local stdio server for desktop clients (MIT — clone and self-host).

hosted endpoint · mcp-remote

{
  "mcpServers": {
    "bountyradar-mcp": {
      "command": "npx",
      "args": [
        "-y", "mcp-remote",
        "https://neeenja--bountyradar-mcp.apify.actor/mcp",
        "--header", "Authorization: Bearer ${APIFY_TOKEN}"
      ],
      "env": { "APIFY_TOKEN": "<your-apify-api-token>" }
    }
  }
}

local · stdio

$ git clone https://github.com/PanStories/bountyradar-mcp
$ cd bountyradar-mcp
$ node src/server.mjs
# then, in your MCP client:
{
  "mcpServers": {
    "bountyradar": {
      "command": "node",
      "args": ["src/server.mjs"]
    }
  }
}

In remote mode the server answers POST /mcp for the protocol. Requests carry your own Apify API token as a Bearer header; self-hosting puts your endpoint at https://<username>--bountyradar-mcp.apify.actor/mcp.

Pricing

Pay per event on the hosted endpoint. No subscription, no seat count, no charge when nothing calls a tool.

$0.0005 · feed / score / stats / subscribe

Charged on each get_bounty_feed, score_bounty, get_stats and subscribe_feed call — half a tenth of a cent each.

Primary billable event
$0.001 · search

Charged on each search_bounties call — a keyword search costs double a feed read because it walks the whole catalog.

Priced per search
Always free

The MCP handshake (initialize), tools/list, the source registry and single-bounty detail lookups cost nothing — an agent can connect, enumerate and inspect at zero cost — and every Apify account carries $5 of monthly platform credit.

Discovery is not metered

At a glance

Search all MCP servers →

Data sources

Opire, BountyHub, ClawHunt, Algora, GitHub issues and validated security programs. The curated catalog ships in-repo; a scraper actor can refresh live records into a separate layer without polluting it. The cost of running the server is compute, not data.
Scores you can audit

Every score is deterministic and explainable: freshness (exponential decay), competition (1/(1+attempts)), agent-solvability (declared or derived from task type) and a reward normaliser, combined into one composite. No black box, no training.
Honeypot guard

Bounties flagged as traps — "reserved for interview loop" and the like — are excluded from the default feed, so an agent does not burn compute or burn its reputation on a decoy.
Intelligence, not a marketplace

Deliberately an intelligence and alerting layer: no fund custody, no KYC, no cross-border settlement. ClawHunt and friends own the settlement layer; this indexes upstream instead of competing.
Transports

Stateless Streamable HTTP for remote hosting on Apify Standby, plus stdio for local desktop clients.
Licence

MIT. Clone, fork and self-host without a commercial agreement; the hosted endpoint is simply the convenient default.