Written and maintained by CASRAI Editorial Board
Published · Last updated
On this page
The Meta Content Library is a controlled-access research platform and programmatic application programming interface (API) that provides qualified academic and non-profit researchers with comprehensive, near-real-time public data from Facebook and Instagram. Launched by Meta Platforms, Inc. to replace CrowdTangle, the library provides access to public posts, pages, groups, events, and creator profiles along with granular engagement metrics and comment metadata. To protect user privacy and fulfill regulatory obligations under frameworks such as the European Union Digital Services Act (Article 40), access is independently vetted and administered through the Inter-university Consortium for Political and Social Research (ICPSR) at the University of Michigan, with computational queries executed inside secure digital cleanrooms.
What it is used for in research
The Meta Content Library serves as the primary data infrastructure for computational social science, media studies, political communication, and public health surveillance across Meta platforms. Following the deprecation and closure of CrowdTangle in August 2024, the academic community required a transparent and methodologically sound pathway to examine digital social dynamics. The Content Library grants vetted investigators systematic access to historical and real-time public content originating from public Facebook Pages, public groups, verified profiles, events, and public Instagram professional and creator accounts.
In computational social science and political research, scholars leverage the platform to track information diffusion, evaluate the virality of competing political narratives, examine electoral communication, and monitor cross-platform coordination during democratic elections. The platform allows researchers to observe not only textual messages but also propagation trajectories of embedded links, multimodal assets, and public re-shares across geographical regions and temporal intervals.
In public health, epidemiology, and crisis communication, investigators use the library to assess health communication campaigns, identify emerging misinformation patterns regarding vaccine efficacy or disease outbreaks, and evaluate community adherence to official guidance. Emergency management researchers analyze real-time disaster response communication, observing how municipal emergency services, humanitarian non-governmental organizations, and affected citizens exchange critical logistical information during extreme weather events and humanitarian crises.
Unlike commercial advertising interfaces or deprecated public scraping scripts, the Meta Content Library structures all observational material as high-integrity research data. Each record is accompanied by comprehensive platform metadata, including precise ISO 8601 creation timestamps, publisher categorizations, verified view counts, and fine-grained emotional reaction breakdowns (likes, loves, wows, hahas, sads, and angry reactions). This structured environment ensures that observational datasets adhere to modern FAIR data principles within controlled computational parameters, supporting replicable scientific analysis without exposing users to privacy violations.
How to install or access it
Access to the Meta Content Library is strictly gated and governed through independent institutional partnerships. It cannot be accessed via unauthenticated public endpoints, open-market API tokens, or commercial advertising developer accounts. Prospective research teams must satisfy rigorous scientific and institutional criteria before access is provisioned.
Institutional Eligibility and Governance
To qualify for access, applicants must hold a demonstrated research appointment at an accredited academic institution or an established non-profit research organization dedicated to scientific or public-interest inquiry. Commercial entities, political organizations, marketing agencies, and independent hobbyists are ineligible. The research proposal must pursue genuine scientific, sociological, or governance questions. Furthermore, projects involving human subjects or observational online behavior must secure an Institutional Review Board (IRB) approval or official formal exemption from their institution ethics board.
Application vetting is conducted independently of Meta commercial divisions. In North America and across global institutions, applications are managed and evaluated by the Inter-university Consortium for Political and Social Research (ICPSR) through its Social Media Archive (SOMAR) portal at the University of Michigan. For researchers operating within the European Union, access is coordinated through CASD (Centre d’Accès Sécurisé aux Données) and partner academic vetting bodies under the framework of Article 40 of the Digital Services Act (DSA).
Onboarding and Legal Agreements
The access process requires executing formal legal instruments between the researcher home university and the governing data repository. The institution must execute a binding data transfer agreement and accept the Meta Research Tools Terms and Conditions. These terms explicitly forbid attempting to re-identify anonymous users, cross-referencing platform identifiers with external personal dossiers, or extracting raw microdata outside the authorized computational workspace.
Execution Environment and Python Client Setup
The Meta Content Library API does not operate on a researcher local laptop or personal workstation. Instead, all programmatic interactions take place inside a secure digital cleanroom: Meta Secure Research Environment (SRE)—hosted on Amazon WorkSpaces virtual desktop infrastructure—or ICPSR Virtual Data Enclave (VDE). These cleanroom workspaces prevent outbound data leakage while providing high-performance computing resources, including JupyterLab, Python 3, and R programming runtimes.
Inside the pre-configured SRE terminal or notebook environment, the official Python library is loaded directly. Researchers initialize the client using the pre-installed package without needing local pip installations:
from metacontentlibraryapi import MetaContentLibraryAPIClient
# Initialize the client connection using the latest production schema
client = MetaContentLibraryAPIClient(version=MetaContentLibraryAPIClient.LATEST_VERSION)
For quantitative researchers working in R, the identical underlying Python client is accessed seamlessly through the R reticulate interface, ensuring consistent query syntax across multi-disciplinary teams.
Step-by-step instructions
Conducting an empirical investigation using the Meta Content Library API involves a structured, sequential workflow. The following step-by-step procedure guides researchers from initial environment connection to data extraction and budget tracking inside the Secure Research Environment (SRE).
- Connect to the Secure Research Environment: Launch the Amazon WorkSpaces client application on your workstation. Authenticate using your institutional credentials and multi-factor authentication (MFA) tokens assigned during the ICPSR onboarding workflow. Once the virtual desktop initializes, open the system terminal or application menu.
- Launch JupyterLab: From the SRE application launcher or terminal, start the pre-configured Jupyter server by entering
jupyter lab. In the web interface, open a new notebook and select the Python 3 kernel. - Import and Initialize the API Client: Execute the initialization cell to establish the secure connection to the Content Library backend:
from metacontentlibraryapi import MetaContentLibraryAPIClient client = MetaContentLibraryAPIClient() client.set_default_version(client.LATEST_VERSION) - Verify Available Query Quotas and Budget: Meta enforces rolling 7-day query quotas per research project. Before constructing large queries, verify your remaining quota using the
/budgetsendpoint:budget_check = client.get(path="budgets") print("Current project budget status:", budget_check.json()) - Construct and Test a Synchronous Preview Query: Preview endpoints allow researchers to validate query filters, keyword parameters, and date bounds with immediate response times. Synchronous preview queries are capped at a maximum return of 1,000 items:
preview_parameters = { "q": "renewable energy policy OR solar subsidies", "since": "2025-01-01", "until": "2025-03-31", "limit": 50, "sort": "newest_to_oldest" } preview_result = client.get(path="facebook/posts/preview", params=preview_parameters) preview_data = preview_result.json() print(f"Preview records returned: {len(preview_data)}") - Dispatch an Asynchronous Batch Job for Large-Scale Ingestion: When your analysis requires thousands or millions of posts exceeding preview thresholds, submit an asynchronous background snapshot job using the
client.post()method:batch_parameters = { "q": "renewable energy policy OR solar subsidies", "since": "2025-01-01", "until": "2025-03-31", "mode": "SNAPSHOT", "name": "renewable_energy_q1_2025_corpus" } job_submission = client.post(path="facebook/posts/job", params=batch_parameters) active_job_id = job_submission.json().get("job_id") print(f"Asynchronous job successfully queued. Job ID: {active_job_id}") - Monitor Job Status and Ingest Results: Poll the job endpoint periodically to verify execution completion. Once the status transitions to
COMPLETED, load the records into a pandas DataFrame for statistical modeling:import time, pandas as pd while True: status_check = client.get(path=f"facebook/posts/job/{active_job_id}") job_status = status_check.json().get("status") print(f"Checking job status: {job_status}") if job_status in ["COMPLETED", "FAILED"]: break time.sleep(30) if job_status == "COMPLETED": dataset_path = status_check.json().get("output_path") df = pd.read_parquet(dataset_path) print(f"Ingested {len(df)} records. Dimensions: {df.shape}")
Output and how to read it
The Meta Content Library API outputs rich, structured JSON and Parquet payloads adhering to the official Meta Content Library Data Dictionary. Understanding the schema architecture and privacy safeguards is vital for accurate computational modeling and causal inference.
Core Schema Entities and Data Fields
Each retrieved record represents a distinct social media publication or institutional entity, organized into specific attribute groupings:
- Publication Identifiers:
idrepresents the unique internal string identifier assigned by the library (e.g.fb_post_...orig_post_...). Thepost_urlprovides the public, canonical web link to the original content on Facebook or Instagram. - Temporal and Textual Attributes:
created_timerecords the exact UTC publication timestamp in ISO 8601 formatting. Themessageorcaptionfield contains the complete textual body authored by the account, whilecontent_typeclassifies the asset (e.g. photo, video, status update, shared link, reel, or album). - Audience Exposure and Engagement Metrics: The Content Library delivers comprehensive exposure indicators. The
view_countattribute supplies verified impression metrics, allowing researchers to calculate true engagement ratios rather than relying solely on engagement counts. Aggregate reaction counts (reaction_count) are partitioned into individual emotional categories:like,love,wow,haha,sad, andangry. In addition,comment_countandshare_counttrack conversational depth and viral dissemination. - Publisher and Account Metadata: Account-level descriptors include
page_id,account_id, accountname, verified badge status (verification_status), registered topicalcategory, and total audience scale indicated byfollower_count. - Multimodal and Visual Features: When queries specify
search_scope="post_text_and_image_text", the response includesimage_text, containing extracted optical character recognition (OCR) transcriptions of text embedded in memes, infographics, and photograph banners.
Privacy Protection and Aggregation Boundaries
The library incorporates strict differential privacy protections to prevent the surveillance of private individuals. Content from personal Facebook profiles or Instagram accounts with fewer than 100 followers is programmatically excluded. Private groups, secret groups, and non-public direct interactions are completely inaccessible. Where individual comments are analyzed, the output presents digitally anonymised data structures stripped of personal identifiers to safeguard citizen privacy while preserving aggregate statistical utility.
Interpreting the Quota and Budget Interface
Calls to the /budgets endpoint return operational quota diagnostics. The payload details allocated credits, records returned within the active 7-day rolling window, and remaining ingestion capacity. Researchers must monitor these metrics to pace batch queries and avoid job rejections during time-sensitive analytical pipelines.
Troubleshooting
Operating within secure cleanroom enclaves and programmatic API constraints introduces operational and technical challenges. Below are common error conditions, root causes, and verified technical remedies.
1. HTTP 400 Bad Request and Parameter Syntax Errors
Symptom: The API returns HTTP status 400 with messages indicating invalid query syntax or unrecognized parameter names.
Cause: Formatting errors in the q search parameter or date bounds. Boolean operators must be strictly capitalized (AND, OR, NOT); lowercase operators (e.g. and) trigger parser exceptions. Additionally, dates must follow the strict four-digit year pattern YYYY-MM-DD or valid integer UNIX timestamps.
Remedy: Verify that search terms are grouped logically with parentheses when mixing boolean operators. Ensure all requested field names match entries in the Meta Content Library Data Dictionary exactly.
2. HTTP 401 Unauthorized and HTTP 403 Forbidden
Symptom: Calls fail with authentication or authorization rejections.
Cause: SRE session timeout, expiration of annual research credentials, or querying an unauthorized platform submodule (e.g. requesting WhatsApp channel tables when the project proposal only specified Facebook Pages).
Remedy: Re-authenticate through the Amazon WorkSpaces access portal. Verify project scope in the Meta Research Tools Manager dashboard. If additional platform modules are required for your study, submit a formal project amendment through the ICPSR SOMAR administrative team.
3. HTTP 429 Rate Limit Exceeded and Rolling Budget Depletion
Symptom: Synchronous or asynchronous requests are rejected with status 429.
Cause: The research team has exhausted its rolling 7-day query or record quota.
Remedy: Execute client.get(path="budgets") to review rolling quota expiration timestamps. Stagger data collection workflows across multiple days or refine keyword filtering parameters to reduce unnecessary record volume. Target specific accounts via page_ids rather than executing unbounded global keyword searches.
4. Synchronous Preview Result Truncation (1,000 Record Ceiling)
Symptom: Queries return exactly 1,000 records despite millions of relevant posts existing on the platform.
Cause: The /preview endpoints are engineered solely for exploratory testing and enforce a hard limit of 1,000 records per call.
Remedy: Transition from client.get() preview calls to asynchronous batch jobs using client.post() on the appropriate /job endpoint with mode="SNAPSHOT".
5. Cleanroom Data Export and Download Restrictions
Symptom: Attempting to download CSV files, raw Parquet tables, or notebook outputs to a local machine fails, or clipboard copy functions are disabled.
Cause: By design, Meta SRE and ICPSR Virtual Data Enclave prohibit outbound raw data transmission to enforce institutional data protection standards.
Remedy: Conduct all statistical transformations, machine learning pipelines, and aggregations inside the cleanroom. When ready to publish, submit aggregate summary tables, regression coefficients, and visualization graphics to the ICPSR disclosure review committee. Approved aggregate outputs are cleared for export to your local workstation.
Free alternatives
While the Meta Content Library provides deep observational data across Facebook and Instagram, its rigorous institutional vetting, strict eligibility rules, and cleanroom operational model may present barriers for smaller projects, rapid pilot studies, or independent scholars. Several open, free, or academic-tier platforms provide alternative social data access for computational research.
The following table summarizes prominent academic and open data interfaces available for digital communication research:
| Platform / Tool | Primary Research Scope | Access Model and Eligibility | Programmatic Interface | Methodological Limitations |
|---|---|---|---|---|
| TikTok Research API | Short-form video trends, virality, youth culture, political discourse | Application required; accredited academic institutions and non-profits only | Direct REST API, developer portal access tokens | Daily query ceilings; historical limits; non-public or deleted videos excluded |
| YouTube Data API v3 | Long-form video metadata, channel statistics, public comments, captions | Open to all registered Google Cloud Console developers (free quota tier) | REST API, official Python and Java client libraries | Daily quota units (10,000 units/day default); complex queries consume quotas rapidly |
| Bluesky AT Protocol Firehose | Decentralized social discourse, microblogging, federated network graphs | Completely open public access; no institutional application or vetting required | WebSocket Firehose, REST API, Python (atproto SDK) | Smaller overall user base than Meta; network topology differs from centralized platforms |
| Reddit Data API (Academic) | Community discussions, subreddit governance, niche topic forums, threaded comments | Developer application required; academic research exceptions available | OAuth2 REST API, Python (PRAW library) | Rate limits strictly enforced; historical archival access curtailed post-2023 policy changes |
| Wikimedia EventStreams and REST API | Information curation, encyclopedia editing dynamics, public attention cycles | Completely open public access; no authentication needed for standard reading | Server-Sent Events (SSE) EventStreams, MediaWiki REST API | Reflects encyclopedic curation rather than conversational social media interaction |
| Internet Archive Wayback Machine | Historical web snapshots, deleted public pages, longitudinal site archives | Completely open public access and Python CDX API | CDX Server API, Python (waybackpy, internetarchive) | Sparse temporal sampling; dynamic client-rendered social feeds often captured incompletely |
For an expanded comparison of regulatory compliance, independent vetting authorities, and computational execution environments across leading social research tools, consult the dataset table below.
Frequently asked questions
How does Meta Content Library differ from CrowdTangle?
CrowdTangle was an industry-focused social monitoring tool that provided public post counts, engagement velocity scores, and CSV downloads across Facebook, Instagram, and Reddit, but it ceased operations in August 2024. In contrast, the Meta Content Library is an academically focused research platform that provides deeper, near-real-time data—including verified view counts, demographic publisher attributes, and optical character recognition for memes. However, unlike CrowdTangle, the Content Library restricts data manipulation to secure virtual enclaves and requires independent institutional vetting via ICPSR.
Who is eligible to apply for Meta Content Library access?
Access is strictly limited to qualified researchers, postdoctoral fellows, and faculty members affiliated with accredited colleges, universities, or non-profit research organizations whose core mission is scientific or public-interest research. Commercial enterprises, political consultants, marketing agencies, and independent researchers without formal academic or non-profit institutional backing are ineligible.
Can researchers export raw data files to their local machines?
No. To prevent personal data exfiltration and protect platform user privacy, researchers cannot download raw microdata, JSON arrays, or post-level CSV files outside the Secure Research Environment (SRE) or ICPSR Virtual Data Enclave (VDE). All computational analysis, natural language processing, and statistical regression modeling must execute within the enclave. Only aggregated statistical summaries, model parameters, and approved visualizations may be exported following formal disclosure review.
What social media platforms are accessible through the library?
The library covers public content across Facebook and Instagram. On Facebook, this includes public Pages, public groups, verified profiles, public events, and public marketplace listings. On Instagram, coverage encompasses business accounts, creator accounts, verified profiles, and public personal accounts that possess 100 or more followers. Monitored research modules also provide selective access to public WhatsApp channels and Threads content.
How does the platform protect individual user privacy?
The Meta Content Library enforces multiple privacy filters designed in collaboration with academic data governance bodies. Private user profiles, closed Facebook groups, direct messages, and accounts with fewer than 100 followers are completely omitted from the data repository. Furthermore, user comments are presented with differential privacy protections or aggregated structures to prevent the re-identification of private citizens participating in public discourse.
Sources
- Meta for Developers: Meta Content Library and API Documentation. https://developers.facebook.com/docs/meta-content-library (accessed 2026-11-16).
- Inter-university Consortium for Political and Social Research (ICPSR): Social Media Archive (SOMAR) Research Platform. https://www.socialmediaarchive.org/ (accessed 2026-11-16).
- Meta Research: Research Tools and Data Access Programs. https://research.facebook.com/ (accessed 2026-11-16).
- European Commission: Digital Services Act (DSA) Article 40 Data Access Framework. https://digital-strategy.ec.europa.eu/en/policies/dsa-data-access (accessed 2026-11-16).
Data: Academic Social Media Research API Parameters and Data Access Requirements
Comparative cross-platform compilation of data access models, vetting authorities, query environments, ingestion quotas, and minimum account thresholds for computational social science and platform research.
Period covered: 2024/2026. Coverage: Global.
| Platform API | Host Entity | Vetting Authority | Execution Environment | Ingestion Quota or Budget | Minimum Follower Threshold |
|---|---|---|---|---|---|
| Meta Content Library API | Meta Platforms, Inc. | ICPSR / CASD Independent Vetting | Secure Research Environment (AWS WorkSpaces / VDE) | Rolling 7-day budget tracked via /budgets | 100+ followers or verified status |
| TikTok Research API | ByteDance Ltd. | TikTok Academic Access Review Board | Developer Portal API (Direct REST / SDK) | Up to 1,000 queries per day per project | Public accounts with public videos |
| YouTube Data API v3 | Google LLC | Google Cloud Console Project Quota | Direct REST Client / Local SDK | 10,000 quota units per day (free default) | None (all public channels and videos) |
| Bluesky AT Protocol API | Bluesky Social, PBC | Open Access / Self-Provisioned Handle | Direct Firehose / WebSocket / REST | Public network rate limits (3,000/5m unauth) | None (federated public data repository) |
| Reddit Data API (Academic) | Reddit, Inc. | Reddit Developer Platform Approval | OAuth2 Direct REST Client | 100 queries per minute with approved token | None (public subreddits and submissions) |
Sources
- Meta Content Library and API Documentation (accessed 2026-11-16)
- ICPSR Social Media Archive (SOMAR) Research Platform (accessed 2026-11-16)
- TikTok Research API Developer Documentation (accessed 2026-11-16)
- Google Developers: YouTube Data API Overview (accessed 2026-11-16)
- AT Protocol Documentation: Bluesky Social Network (accessed 2026-11-16)
Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.







