How to Get Twitter Data Without the Official API (September 2026)

As of September 9, 2026, the official X API has no free tier and no academic tier: new developers pay per use ($0.005 per post read, 7-day search only, a cap of 3 million post reads a month), and full-archive search back to 2006 is sold only through Enterprise contracts that third-party estimates put above $42,000 a month. The working routes around it are third-party data APIs (TwitterAPI.io from $0.15 per 1,000 tweets, SocialData.tools from $0.20, Sorsa from $49 a month), licensed multi-platform platforms you can query from an AI assistant over MCP, open-source scrapers that log in with your own X accounts (twscrape, Twikit), and static datasets. Nitter went offline on August 25, 2026, and snscrape has not worked for X since 2023.

This page is a dated status guide. The date in the title is when every route below was last checked against the provider's own pricing page or the project's repository, and we refresh it when something changes. Legacy tier prices still quoted elsewhere (Basic at $100 or $200 a month, Pro at $5,000) are closed to new signups.

Disclosure: we build xpoz, a licensed social data platform that appears under Method 2 and in the "How Xpoz Addresses This" section. The other routes are listed on their own merits, with facts taken from their own pages at publication time.

What Still Works in September 2026

The Official API in September 2026

X moved every new developer to pay-per-use on February 6, 2026. There is no subscription: you load credits into the developer console and each request deducts a per-resource fee, $0.005 for a post read, $0.010 for a user profile, $0.001 for reads of your own data. Pay-per-use covers the last 7 days of search, has no streaming endpoints, and caps post reads at 3 million per monthly billing cycle before an Enterprise contract is required.

The legacy tiers are closed. Basic ($200 a month at the end) stopped accepting signups and its remaining subscribers have been migrated to pay-per-use since June 2026. Pro ($5,000 a month), the only self-serve tier that ever included full-archive search, is closed permanently. Enterprise is the sole official route to the full archive, the firehose and engagement automation, priced by negotiation at $42,000 a month and up according to third-party pricing pages. X publishes the current per-request fees on its pricing page.

For a complete breakdown of what each operation costs and when alternatives make more sense, see our Twitter API Pricing Guide.

For many legitimate use cases (academic research, market analysis, small business intelligence, startup MVPs) these prices create barriers. Reading 100,000 tweets with author profiles costs $1,500 a month at pay-per-use rates, and any question that reaches further back than 7 days has no self-serve answer at all. This gap is why the alternatives below exist.

What You Can Legally Access

Twitter/X data falls into several categories:

Public data: Tweets, profiles, follower lists, and engagement metrics that any logged-in user can see. This data is generally accessible through various methods.

Protected data: Content from private accounts, direct messages, and account-level analytics. This requires explicit authorization and typically only the official API with proper OAuth flows.

Rate-limited data: Some queries (like full-text search across historical archives) are restricted regardless of method.

The alternatives discussed here focus on public data access, the same information visible to any Twitter user browsing the platform.

Method 1: Third-Party Data APIs

The most straightforward alternative to the official API is using third-party services that provide Twitter data through their own APIs or interfaces.

How They Work

Third-party providers maintain their own data infrastructure:

1. Data collection: They gather public Twitter data through various methods
2. Storage and indexing: Data is stored, indexed, and made searchable
3. API access: Customers query data through REST APIs, MCP integrations, or web interfaces

Advantages

• Lower cost: The official API charges $5.00 per 1,000 post reads; the providers below charge $0.15-$0.40 per 1,000 tweets
• No 15-minute windows: Most providers bill per call or per month without per-window rate limits
• Historical access: Several providers reach further back than the official API's 7-day search
• Multi-platform: Some providers cover Instagram, TikTok, Reddit alongside Twitter

Considerations

• Data freshness: Some lag between real-time and when data appears (usually minutes to hours)
• Coverage gaps: May not capture 100% of tweets, especially from low-engagement accounts
• Terms of service: Operating in a legal gray area depending on jurisdiction and use case

Providers Working in September 2026

When evaluating third-party data APIs, consider the pricing model (per-request, subscription, or pay-as-you-go), how far back the index reaches and what metadata it includes, query capabilities (full-text search, boolean operators, filters), output formats, and any residual rate limits. Our Best Twitter API Alternatives in 2026 comparison covers more providers side by side.

Method 2: AI-Native Data Access (MCP)

A newer approach leverages the Model Context Protocol (MCP) to access social media data through AI assistants like Claude or ChatGPT, and through coding agents like Claude Code and Cursor.

How It Works

Instead of writing code against a REST API, you interact with social data through natural language:

"Find tweets about 'climate change' from the past week
with more than 1000 retweets"

The MCP server translates this into appropriate data queries and returns structured results. The same server answers a coding agent's tool calls, so a Claude Code session can pull tweets into a script without an API integration step.

Advantages

• No coding required: Natural language interface eliminates API integration work
• Exploratory analysis: Easy to iterate on queries without rewriting code
• Integrated workflow: Data access within the same interface used for analysis
• Multi-step operations: AI can combine multiple queries automatically

Considerations

• Requires an MCP client: Works through Claude, ChatGPT, Claude Code, Cursor or an agent framework
• Less programmatic: Not ideal for automated pipelines on its own (though the same platforms usually expose REST APIs and SDKs for that)
• Learning curve: Understanding effective prompting takes practice

Best For

• Researchers doing exploratory analysis
• Teams already using AI assistants in their workflow
• Non-technical users who need data access without coding
• Rapid prototyping before building dedicated integrations

Method 3: Web Scraping

Building custom scrapers, or running open-source ones, to extract Twitter data directly from the web interface.

How It Works

Scrapers use browser automation or HTTP requests to:

1. Navigate to Twitter pages (profiles, search results, threads)
2. Extract data from the rendered HTML or underlying API calls
3. Parse and store the results

Which Twitter Scrapers Still Work (September 2026)

Every working X scraper in September 2026 logs in as a user. That is the practical difference from a year ago: the logged-out routes (Nitter mirrors, snscrape's public endpoints) are gone, so scraping now means supplying your own accounts and accepting that they can be rate-limited or banned. For the status of scrapers on Reddit, Instagram and TikTok too, see Open-Source Social Media Scrapers: What Still Works in September 2026.

Advantages

• Full control: Build exactly what you need
• No third-party dependencies: Direct access to source
• One-time cost: Build once, run indefinitely (in theory)

Significant Challenges

• Maintenance burden: Twitter frequently changes its frontend, breaking scrapers
• Anti-bot measures: Rate limiting, CAPTCHAs, account suspensions
• Legal risk: Potential terms of service violations; X Corp's 2026 action against Nitter cited the Texas Harmful Access by Computer Act and the Lanham Act
• Technical complexity: Requires ongoing engineering investment
• Scale limitations: Difficult to achieve high-volume reliable extraction

Reality Check

While scraping seems attractive (it's "free"), the hidden costs are substantial:

• Engineering time to build and maintain scrapers
• Infrastructure costs for proxies, compute and spare accounts
• Risk of account bans affecting your business
• Unreliable data quality and coverage

For most use cases, third-party APIs are more cost-effective when accounting for total engineering investment.

Method 4: Academic and Research Programs

X retired the Academic Research track in 2023, and as of September 2026 there is no academic tier in the X API product line. X's own research guidance points researchers at the same pay-per-use and Enterprise access as everyone else, and threads on the X developer forum asking how to apply for academic access have gone unanswered since the program closed.

What researchers use instead in 2026:

• Pay-per-use for small, recent samples: $0.005 per post read covers a few thousand tweets from the last 7 days for under $20
• Third-party data APIs for larger or older samples: TwitterAPI.io states full-history access at $0.15 per 1,000 tweets
• Licensed platforms for cross-platform studies where Twitter is one of several sources
• Static datasets for reproducible corpora with a fixed time range (Method 5)

Institutional review boards and journals increasingly ask how data was obtained. Whichever route you take, record the provider, the date range and the collection date, since coverage differs between sources and changes over time.

Method 5: Data Marketplaces and Datasets

Pre-collected Twitter datasets available for purchase or download.

Sources

• Academic datasets: Research institutions sometimes share collected data
• Data marketplaces: Companies sell pre-packaged Twitter datasets
• Open datasets: Some historical Twitter data is publicly available

Best For

• Historical analysis where real-time data isn't needed
• Training machine learning models
• Research requiring specific time periods

Limitations

• Static snapshots, no real-time updates
• May not include the specific data you need
• Quality and coverage varies significantly
• Licensing restrictions on use

Choosing the Right Approach

Decision Framework

Your use case determines the best method:

Your technical resources matter:

• No developers: MCP-based access or no-code platforms
• Small team: Third-party APIs with good documentation
• Engineering capacity: Any method, evaluate total cost of ownership

Your budget constraints:

• Minimal budget: Free tiers and trial credits, or scrapers with your own accounts
• Moderate budget: Third-party API subscriptions or per-call billing
• Larger budget: Official pay-per-use for recent data, Enterprise for the full archive

How Xpoz Addresses This

Xpoz provides Twitter data access through the Model Context Protocol, enabling natural language queries through AI assistants and tool calls from coding agents, alongside a REST API, SDKs and a CLI for programmatic use.

Connect From Claude Code or Claude Desktop

From Claude Code, one command adds the remote server; there is nothing to install locally and authentication is a Google sign-in on first use:

claude mcp add --transport http --scope user xpoz https://mcp.xpoz.ai/mcp

In Claude Desktop, add a custom connector under Settings, Connectors, with the URL https://mcp.xpoz.ai/mcp. ChatGPT, Cursor and agent frameworks connect to the same URL.

Core Capabilities

Search tweets by keyword:

Tool: getTwitterPostsByKeywords
- query: "artificial intelligence" AND startups
- fields: ["id", "text", "authorUsername", "retweetCount", "likeCount"]

Get user profiles and connections:

Tool: getTwitterUserConnections
- username: "target_account"
- connectionType: "followers"
- fields: ["username", "followersCount", "description"]

Analyze engagement:

Tool: getTwitterPostInteractingUsers
- postId: "tweet_id"
- interactionType: "commenters"

Track mention volumes:

Tool: countTweets
- phrase: "brand name"
- startDate: "2026-08-09"
- endDate: "2026-09-09"

Practical Workflow Example

Suppose you're researching sentiment around a product launch:

1. Find relevant tweets: "Search for tweets mentioning 'ProductX launch' from the past 2 weeks"

2. Identify key voices: "Who are the most-followed users discussing this topic?"

3. Analyze engagement: "Which tweets got the most replies and quotes?"

4. Export for analysis: "Export these results to CSV for further analysis"

Each step uses natural language, with the MCP server handling the underlying data operations.

Coverage

Xpoz indexes billions of posts across Twitter, Instagram, TikTok, and Reddit. For Twitter specifically:

• Real-time plus historical data, with rolling 60-day windows
• Tracked keywords and accounts, which get better coverage and fresher results
• User profiles and metadata
• Follower/following relationships
• Engagement metrics (likes, retweets, replies, quotes)

Pricing

• Free tier: up to 75,000 results (one-time)
• Pro ($20/month): up to 1 million results/month
• Max ($200/month): up to 18 million results/month

No API keys to manage, no rate limit errors to handle, no code to write.

Implementation Considerations

Data Quality

Regardless of method, consider data quality:

• Completeness: Does the source capture all relevant tweets?
• Accuracy: Is metadata (timestamps, metrics) reliable?
• Freshness: How quickly does new content appear?
• Authenticity: Can you filter out bot accounts?

Compliance

Consider legal and ethical aspects:

• Terms of service: Understand what's permitted
• Privacy regulations: GDPR, CCPA implications for personal data
• Research ethics: IRB approval if applicable
• Data retention: How long can you store collected data?

Scalability

Think about growth:

• Volume increases: Will your method scale with your needs?
• Cost scaling: How does pricing change at higher volumes?
• Performance: Query speed at larger data sizes?

Key Takeaways

• The official API is pay-per-use only: No free tier, no academic tier, 7-day search, and the full archive behind Enterprise contracts since February 2026.

• Third-party APIs offer the best balance for most use cases: $0.15-$0.40 per 1,000 tweets, maintained infrastructure, and in some cases full-history access.

• MCP-based access is now mainstream: Claude, ChatGPT and coding agents such as Claude Code can query social data directly, without an integration step.

• Logged-out scraping is over: Nitter is offline and snscrape no longer works for X; twscrape and Twikit work only with your own accounts.

• Match method to use case: Academic researchers, startups, and enterprises have different optimal paths.

• Consider total cost of ownership: Factor in development time, maintenance, and reliability, not just subscription fees.

Frequently Asked Questions

Can I still get Twitter data without the official API in 2026?

Yes. As of September 2026 the working routes are third-party data APIs (TwitterAPI.io, SocialData.tools, Sorsa), licensed platforms with MCP and REST access such as xpoz, open-source scrapers that log in with your own accounts (twscrape, Twikit), and static datasets. Logged-out routes such as Nitter and snscrape no longer work.

Is there a free Twitter API in September 2026?

Not an official one. X discontinued the free tier when it moved to pay-per-use on February 6, 2026. Free ways to start: TwitterAPI.io's $0.10 signup credit, SocialData.tools' 3 free requests a minute, xpoz's one-time free tier covering up to 75,000 results, and twscrape with accounts you already own.

Is snscrape still working for Twitter?

No. The repository is not archived, but its last commit is from November 2023 and its Twitter modules depend on public endpoints that X removed when it put search behind a login wall in 2023. For X, the maintained open-source options are twscrape and Twikit, both of which need your own accounts.

Is Nitter still available?

No. X Corp sent the Nitter project a cease-and-desist letter dated August 24, 2026 demanding a takedown of every instance and the repository. nitter.net went offline on August 25, the GitHub repository is archived, and XCancel and the RSS bridges built on Nitter went down with it.

Does X still offer an Academic Research API?

No. The Academic Research track stopped accepting applicants in 2023 and X's product line in 2026 has no academic tier. Researchers use pay-per-use for small recent samples, third-party data APIs or licensed platforms for larger or older ones, and static datasets for reproducible corpora.

How do I get tweets older than 7 days without an Enterprise contract?

Pay-per-use search covers only the last 7 days. Older data comes from third-party providers (TwitterAPI.io states full-history access at $0.15 per 1,000 tweets), licensed platforms, or datasets with a fixed time range. Enterprise, at $42,000 a month and up, is the only official path to the full archive.

Conclusion

Getting Twitter data without the official API is not only possible but often preferable for many use cases. The ecosystem of alternatives has matured, offering options from per-call REST APIs to AI-native natural language interfaces, while the free logged-out scraping stack of earlier years has largely disappeared.

For most users, the practical path is:

1. Start with MCP-based access if you're already using AI assistants or coding agents: it's the fastest path to exploratory insights
2. Move to third-party APIs when you need programmatic access or higher volumes
3. Consider the official API only when you need write access or specific compliance requirements

The key is matching your method to your actual needs. A researcher analyzing a few thousand tweets has different requirements than a company building real-time monitoring infrastructure. Choose accordingly, and you'll find that Twitter data is more accessible than the official pricing might suggest.

Related Reading

• Open-Source Social Media Scrapers: What Still Works in September 2026
• Understanding Twitter API Pricing Tiers and Alternatives
• Best Twitter API Alternatives in 2026
• Twitter/X Data Access Options Compared (2026)
• How to Export Twitter/X Data to CSV with an AI Agent