Using AgentFlare Social, Predictions and Crypto Data

This guide is for developers and agent builders who want to pull structured data from three AgentFlare directories: Social Pulse, Predictions and Crypto. You will learn which .data.json files exist, what every field in them holds, what each request costs, and how to fetch a file with the X-Crawler-Token header. You will also read prediction events and coin data in Python, and see how the pricing block inside each response is laid out. The examples use only paths and field names that AgentFlare documents.

How the flow works

Before you start

• A crawler token to put in the X-Crawler-Token header
• curl and Python 3 installed
• Basic familiarity with reading JSON

Step by step

1. Know which files are paid

   Index pages, .html articles and .md teasers are free (tier L1_free). Every *.data.json and *.full.md file under /social/, /predictions/ and /crypto/ is paid (tier L2_paid) and metered per request at 0.01 USD each. That includes the index.data.json file in each directory. For comparison, the news directory is 0.005 and research and scholar are 0.02. Plan your call count before looping over many files.

   S5 · agentflare.org

2. Fetch the Social Pulse index

   Request https://agentflare.org/social/index.data.json and send your token in X-Crawler-Token. This header replaces the earlier name X-AISA-Crawler-Token, which older AISA pages still show. Never put the token in Authorization: Bearer, because that header is reserved for the gateway layer. The response holds data.categories[], where each category has id, name, headline, url and a stats object with posts, likes, views, medLikes and verified.

   curl -s 'https://agentflare.org/social/index.data.json' \
     -H 'X-Crawler-Token: YOUR_CRAWLER_TOKEN'

   S1 · esa.aisa.oneS2 · agentflare.orgS5 · agentflare.org

3. Read prediction events in Python

   The predictions index returns data.count, data.total_volume_usd and data.events[]. Each event carries title, slug, leading, probability, change_pp_30d and volume_usd. The slug points to the per-event file at https://agentflare.org/predictions/<slug>.data.json. Each request is billed at 0.01 USD, so fetch the index once and then only the events you need.

   import json, urllib.request
   
   url = 'https://agentflare.org/predictions/index.data.json'
   req = urllib.request.Request(url, headers={'X-Crawler-Token': 'YOUR_CRAWLER_TOKEN'})
   with urllib.request.urlopen(req) as r:
       doc = json.load(r)
   
   print(doc['data']['count'], doc['data']['total_volume_usd'])
   for e in doc['data']['events']:
       print(e['slug'], e['leading'], e['probability'],
             e['change_pp_30d'], e['volume_usd'])

   S1 · esa.aisa.oneS3 · agentflare.orgS5 · agentflare.org

4. Fetch one event and one coin

   An event file such as bitcoin-above-on-october-10-2026.data.json has data.event, data.slug, data.volume_usd, data.volume_24h_usd, data.end_date, data.markets[] (question, odds, volume_usd) and data.tags. Crypto offers a daily file at YYYY-MM-DD.data.json and per-coin files such as bitcoin.data.json or zcash.data.json. Replace the date with a real date.

   curl -s 'https://agentflare.org/predictions/bitcoin-above-on-october-10-2026.data.json' \
     -H 'X-Crawler-Token: YOUR_CRAWLER_TOKEN'
   
   curl -s 'https://agentflare.org/crypto/zcash.data.json' \
     -H 'X-Crawler-Token: YOUR_CRAWLER_TOKEN'

   S1 · esa.aisa.oneS3 · agentflare.orgS4 · agentflare.org

5. Parse crypto fields and the pricing block

   The crypto index and daily files give data.total_market_cap_usd, data.coin_count, data.advancers_24h, data.stablecoin_share_pct, data.top[] and data.trending[]. A per-coin file adds data.price_usd, data.market_cap_rank, data.fdv_usd, data.change_pct (h24, d7, d30, d90), data.ath_usd, data.atl_usd, data.supply and data.history_90d[]. Every response also ends with a pricing object, which the snippet prints.

   import json, urllib.request
   
   url = 'https://agentflare.org/crypto/zcash.data.json'
   req = urllib.request.Request(url, headers={'X-Crawler-Token': 'YOUR_CRAWLER_TOKEN'})
   with urllib.request.urlopen(req) as r:
       doc = json.load(r)
   
   d = doc['data']
   print(d['symbol'], d['price_usd'], d['market_cap_rank'])
   print(d['change_pct']['h24'], d['change_pct']['d7'])
   print(d['supply']['circulating'], d['supply']['max'])
   for p in d['history_90d'][:3]:
       print(p['date'], p['price'])
   print(doc['pricing']['price_usd'], doc['pricing']['method'])

   S1 · esa.aisa.oneS4 · agentflare.orgS5 · agentflare.org

Social, predictions and crypto files

Common pitfalls

• Do not send the crawler token in Authorization: Bearer. That header is reserved for the gateway layer; use X-Crawler-Token instead.
• Only the .html and .md teasers are free. Every .data.json request, including index.data.json, is metered at 0.01 USD in these three directories.
• Change fields are named differently: the crypto index uses data.top[].change_24h and change_7d, while a coin file nests them as data.change_pct.h24 and d7.
• Replace YYYY-MM-DD in file names like atp-fokina-tiafoe-YYYY-MM-DD.data.json with a real date, or the request will not match a file.

Key takeaways

• Social, predictions and crypto .data.json files each cost 0.01 USD per request; teasers and index pages in .html or .md are free.
• Authenticate with the X-Crawler-Token header, never with Authorization: Bearer.
• Fetch each directory's index.data.json first, then follow its slugs to per-item files.

FAQ

How this page is made. Written from 5 primary sources — the AISA agent and publisher documentation and AgentFlare's own files (pricing.json, robots.txt, llms.txt, live responses, .data.json field lists). Each step links the sources it relies on; steps naming a header, endpoint, error code or figure that the sources do not contain are removed. Refreshed weekly.