Public social platforms
Collect the public conversation your Live Searches define across X, Instagram, TikTok, Facebook, LinkedIn, Snapchat and Reddit.
Social intelligence for research and OSINT
Research the public conversation without losing the evidence.
Build a live, query-led corpus across six public platforms. Trace narrative change through source-linked Mentions, AI enrichment, structured filters, exports, and an open API.
Public social data. Platform-aware collection. Month-to-month access.
Trusted by
A research surface built to be interrogated
Language filters
Filter supported Mentions by language, then retain the original source context for expert review.
Buildable research access
Use plan-gated, scoped API keys to query Mentions, campaigns, Live Searches, analytics, and research outputs.
From research question to inspectable corpus
A defensible finding starts with a defensible corpus.
Keep the research question, collection logic, underlying Mentions, interpretation, and output connected from the first query onward.
Frame the corpus around a real question
Organize work in campaigns, then define Live Searches around actors, organizations, places, events, claims, and the language used by the communities in scope. Query handling adapts to the supported platform rather than pretending every source behaves the same way.
- Campaign-level research organization
- Platform-aware query handling
- Named entity and topic extraction
- Live collection tied to the research scope
Preserve the source behind the interpretation
Move from a chart back to the captured Mentions that produced it. Review the author, source link, timestamp, engagement, media, and enrichment context before turning a pattern into a finding.
- Source-linked Mention review
- Chronological and engagement sorting
- Media, author, and enrichment context
- Review and escalation states
Compare narrative structure, not only volume
Examine sentiment intensity, emotion, intent, entities, topics, location, platform, and engagement together. Use filters and analytics to test where a narrative differs by source, audience, place, or period.
- Sentiment and emotion analysis
- Intent and classification filters
- Entity, topic, and location context
- 40+ analytics views
Move findings into a reproducible workflow
Export CSV, JSON, Excel, or PDF for analysis and review. Use scoped API access and webhooks on eligible plans when the corpus or aggregate results need to flow into notebooks, archives, or reporting systems.
- Structured data exports
- Scoped API keys
- Aggregated analytics endpoints
- Webhook delivery for supported events
A defensible research rhythm
Anatomy of an evidence-led research cycle
Keep observation, interpretation, and output connected as the public conversation changes.
Define
Write the question, actors, source scope, language, and known ambiguity before collection begins.
Collect
Run platform-aware Live Searches and let automatic enrichment structure each captured Mention.
Interrogate
Filter, compare, and return to source material when a pattern appears significant.
Publish
Export or query the evidence and state the scope, limits, and date of the analysis.
Built for scrutiny
The context behind every finding
AI reduces the first pass. It does not replace source criticism, language expertise, methodological judgment, or verification.
Automatic enrichment, visible judgment
Every captured Mention receives structured AI analysis. Researchers can still inspect the underlying content and decide whether the model output belongs in the analysis.
Multilingual and geographic context
Filter across 12 supported languages and use inferred location alongside explicit geographic references. Treat both as analytical signals to verify, not ground truth.
Noise controls before the chart
Relevance analysis, spam filtering, duplicate handling, and search filters help reduce contamination before aggregate patterns reach a briefing or publication.
A shared record for collaborators
Campaigns, statuses, alerts, digests, exports, and API access give researchers, editors, and analysts a common body of source-linked material.
Field-tested, then published
We used Socialhose to build a dated record of Sudan's public conflict conversation
The Sudan Conversation Report (Edition 01) coded 1,179 events from public posts between April and June 2026. The analysis measured the Framing Gap between how violence was narrated and where harm landed, then published its method, period, and findings.
A frozen research corpus collected through the Socialhose Public API and interpreted for the published edition.
Narrative named civilians as targets in 18% of violence while the impact reading put civilian exposure at 71%.
The share of civilian-directed violence whose framing did not name civilians as its object.
Published by Social Knowing, our independent research journal. Report figures are frozen at publication; the live board carries the current reading. Both are built from Socialhose data.
Method, coverage, and access
Questions research teams ask first
Which sources can a standard listening campaign collect?
Live collection supports X, Instagram, TikTok, Facebook, LinkedIn, and Reddit. Query syntax and available public content vary by platform, so the corpus should state its source scope.
How much historical data can we retrieve?
Historical availability and lookback vary by platform and Live Search type. Socialhose is built around live collection and adds historical context where the selected source supports it.
Can we inspect the source behind an aggregate result?
Yes. The workspace lets you move from analytics to the captured Mentions and review source, author, time, media, engagement, and enrichment context. External source availability can change after collection.
Does AI analysis replace human coding or verification?
No. Automatic enrichment structures the first pass across sentiment, emotion, intent, entities, topics, and more. Research teams remain responsible for methodology, source criticism, language context, verification, and published conclusions.
Can we export data or work through an API?
Exports include CSV, JSON, Excel, and PDF. Eligible plans add scoped API keys, webhooks, and aggregated analytics endpoints for buildable research workflows.
Does Socialhose certify research ethics or compliance?
No. Socialhose provides public-conversation collection and analysis tools. Your institution is responsible for ethics review, legal basis, platform terms, data governance, and publication standards for each project.
Start with the question that cannot wait
Build the record before the narrative moves on.
Define the conversation, inspect the context, and carry the evidence into the next stage of your work.