Curated Paid API Directory for AI Agents

Bitcoin market intelligence API with L402 pay-per-call auth. 8 endpoints: price/RSI/MACD, mempool fees, on-chain metrics, derivatives, AI news sentiment, 9-factor BUY/SELL/HOLD signal and full market summary. Designed for developers and autonomous AI agents — includes bot_ready object for zero-parsing consumption. No registration, no API key. 2-200 sats per call.

Timestamp a file on Bitcoin by uploading it directly. File is hashed server-side, stored on Google Cloud, and anchored via OP_RETURN. POST {action: l402_timestamp, file_data: <base64>, filename: <name>} to certvera.com/api/l402. 30,000 sats via L402 (includes 5,000 sat upload surcharge). Max 25 MB. Supports PDF, images, and documents.

Update webhook and email notification settings on an existing Bitcoin timestamp. POST {action: l402_timestamp_update, timestamp_id: <id>, webhook_url: <url>, notification_email: <email>} to certvera.com/api/l402. 10 sats via L402. Modify notification endpoints after initial timestamp submission.

Check the status of a pending or confirmed Bitcoin timestamp. Returns confirmation count, block height, and batch info. POST {action: l402_timestamp_status, timestamp_id: <id>} or {action: l402_timestamp_status, hash: <sha256hex>} to certvera.com/api/l402. Free, rate-limited to 10 req/min. No account needed.

Verify a SHA-256 document hash on the Bitcoin blockchain. Returns OP_RETURN proof with block height, tx ID, and merkle path. POST {action: l402_timestamp_verify, hash: <sha256hex>} to certvera.com/api/l402. Free, rate-limited to 10 req/min. No account needed.

Timestamp documents on Bitcoin via OP_RETURN. 5 actions at POST certvera.com/api/l402: (1) l402_timestamp - anchor a SHA-256 hash, 25k sats L402. (2) l402_timestamp with file_data - upload and hash a file, 30k sats L402. (3) l402_timestamp_verify - check if hash exists on-chain, free. (4) l402_timestamp_status - check pending/confirmed status, free. (5) l402_timestamp_update - subscribe to webhook/email notifications, 10 sats L402. Immutable proof of existence, no account needed.

[Prerequisites] 1. **OpenNode Account**: Sign up at

Machine-readable API specification in OpenAPI 3.0.3 format.

Returns current L402 paywall metadata including free tier and priced endpoints.

Evaluates the quality of who a pubkey follows, scoring each follow by trust score and categorizing into tiers (strong/moderate/weak/unknown). Returns an overall quality score (0-100), breakdown metrics (avg trust, reciprocity, diversity, signal ratio), category counts, and suggestions for low-quality follows to reconsider. Useful for Nostr clients to help users curate healthier follow lists.

Compares the trust circles (mutual follows) of two pubkeys. Returns overlapping members (in both circles), unique members (in only one), and a compatibility score (0-100) based on circle overlap ratio, shared follow ratio, and average WoT score of overlapping members. Useful for Nostr clients to show 'how compatible are these two users?' or 'who do we both trust?'

Returns the trust circle (mutual follows) for a pubkey with per-member scoring, shared follow counts, mutual strength metrics, and aggregate circle analytics including cohesion, density, and role distribution. The inner circle highlights the top 10 most-trusted mutual connections.

WebSocket endpoint for live score updates. Connect, subscribe to pubkeys, receive current scores immediately then updates after each graph recomputation (~6h). Protocol: send {type:subscribe,pubkeys:[...]} to watch up to 100 pubkeys. Without WebSocket upgrade, returns endpoint documentation as JSON.

Returns trust scores for a pubkey from our engine and all known external NIP-85 providers. Includes consensus metrics (mean, median, standard deviation, agreement level). Demonstrates NIP-85 interoperability — different providers independently scoring the same pubkey.

Analyzes up to 50 pubkeys in a single request, returning each one's trust score, percentile rank, follower metrics, mutual connections, 2-hop reach estimate, and network role classification (hub, authority, connector, consumer, observer, participant, isolated). Results sorted by trust score descending. No simulation — uses pre-computed PageRank for fast O(1) per-pubkey lookups.

Performs differential PageRank analysis: computes current scores vs. hypothetical scores after a simulated graph change. Shows which pubkeys would be most affected and by how much. Useful for understanding the cascading impact of a single follow/unfollow on the trust network.

Uses five graph-theoretic link prediction signals (Common Neighbors, Adamic-Adar Index, Preferential Attachment, Jaccard Coefficient, WoT Score Proximity) to estimate the likelihood of a follow relationship forming. Returns a prediction score (0-1), confidence, classification, per-signal breakdown, and top mutual connections.

Computes a composite reputation score (0-100, grade A-F) by combining five dimensions: WoT standing (PageRank percentile), Sybil resistance (follower quality and mutual trust), community integration (cluster membership and quality), anomaly cleanliness (absence of trust manipulation flags), and network diversity (follower spread across graph regions). Returns a detailed breakdown with per-component scores, grades, and a human-readable summary.

Scores multiple pubkeys for Sybil resistance in one request. Uses a simplified scoring model for performance. Results sorted by sybil_score ascending (most suspicious first). Useful for relay operators filtering event streams.

Computes a Sybil resistance score (0-100) by analyzing five signals: follower quality (average WoT score of followers), mutual trust (organic bidirectional relationships), score-rank consistency (PageRank vs follower count alignment), follower diversity (neighborhood spread), and account substance (overall activity). Returns a classification (genuine, likely_genuine, suspicious, likely_sybil), confidence level, and full signal breakdown. Designed for relay operators to gate access or filter cont