Back to blog
Crypto Basics

Pump and Dump Schemes - How to Spot and Avoid Investment Scams

Learn everything about pump and dump schemes and how to identify it, in this descriptive guide.
Token Metrics Team
7 Minutes
Want Smarter Crypto Picks—Free?
See unbiased Token Metrics Ratings for BTC, ETH, and top alts.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
 No credit card | 1-click unsubscribe

As the world of finance continues to evolve, new and innovative investment opportunities have emerged. However, alongside these opportunities come potential risks and scams.

One such scam that has become increasingly prevalent is the "pump and dump" scheme. In this article, we'll discuss what a pump and dump scheme is, how to spot it, and how to avoid falling victim to such investment scams.

What is Pump and Dump?

Pump and dump is a type of investment fraud used in the financial market where one or more people intentionally raise the price of a stock, crypto or other financial assets by disseminating false or deceptive information to lure investors. 

Once the price has increased to the desired level, these manipulators sell their shares at a profit, leading to a decrease in price, and leaving innocent investors with securities that are either worthless or overpriced.

How Does the Pump and Dump Scheme Work?

The mechanics of a pump and dump scheme works by manipulating microcap assets, such as penny stocks or cheap cryptos, which are companies with a small market capitalization. These assets are usually traded at a very low price, and do not follow strict requirements for public listing.

Therefore, scammers can easily manipulate information about these assets. The lack of publicly available information provides fraudsters with favorable conditions since potential investors lack the necessary sources to verify all the information about a company and management.

Moreover, microcap assets are highly illiquid and have very low trading volume. As a result, even small transactions can significantly boost the asset price.

The fraudster typically starts by buying a large number of shares in that type of micro cap asset, then spreading false information about the company through various channels, such as social media, chat rooms, and email newsletters.

Once the price of the asset has been artificially inflated, the fraudsters will begin selling their holdings, realizing significant profits. As the scheme is uncovered, the price of the asset typically plummets, leaving unsuspecting investors with significant losses.

Types of Pump and Dump Schemes

There are several types of pump and dump schemes in financial world, including:

Pre-arranged pump and dump: A group of fraudsters will agree to buy a particular asset simultaneously, creating an artificial demand and price surge. Once the price has reached a certain level, the fraudsters will sell their holdings, causing the price to crash.

Social media pump and dump: Fraudsters will use social media platforms to spread false information and create hype around a particular cryptocurrency, stocks or any other asset, encouraging investors to buy. Once the price has been inflated, the fraudsters will sell their holdings.

Exchange-based pump and dump: Fraudsters will use exchanges with low liquidity to create artificial demand for a particular asset, causing the price to surge. Once the fraudsters have sold their holdings, the price typically crashes.

Role of Social Media in Pump and Dump Schemes

Social media plays a significant role in pump and dump schemes as it allows individuals to disseminate false or misleading information about a stock or cryptocurrency quickly and easily to a large audience. 

Scammers may create fake social media profiles or use existing ones to promote the cryptocurrency or digital asset and spread positive news about it, encouraging others to invest. They may also use social media to organize groups of investors to buy and sell the asset at specific times to manipulate its price. 

Social media platforms such as Twitter and Telegram are popular channels for these activities. Investors must be cautious and verify the accuracy of information before making investment decisions based on social media posts or messages.

Pump and Dump Examples

Here are some real-life examples of crypto pump and dump schemes:

Verge - In April 2018, the cryptocurrency Verge experienced a sudden price increase of over 300% in just a few hours. The price increase was caused by a group of individuals who coordinated a pump and dump scheme on social media platforms. The group created false hype around the cryptocurrency, which caused a massive influx of investors who were looking to get rich quickly. 

Soon after the scheme was executed, the value of the cryptocurrency dropped sharply, resulting in substantial losses for investors who failed to dispose of their holdings promptly.

DogecoinDogecoin is a cryptocurrency that was created as a joke in 2013 but gained a significant following due to its popularity on social media. In 2021, a group of individuals on Reddit coordinated a pump and dump scheme on Dogecoin, causing its value to increase significantly in a very short span of time.

The group created a social media campaign that encouraged investors to buy and hold Dogecoin, which caused its value to skyrocket. However, as with all pump and dump schemes, the value of the cryptocurrency eventually plummeted, causing significant losses for investors who were not able to sell their holdings in time.

How to Spot Pump and Dump?

There are several warning signs that investors can look for to spot a pump and dump scheme -

High-pressure sales tactics - Fraudsters often use high-pressure sales tactics to convince investors to buy into a stock or crypto quickly before the price goes up. They may use language such as "this is a once-in-a-lifetime opportunity" or "you don't want to miss out on this."

Unsolicited investment advice - Be wary of unsolicited investment advice from people you don't know, especially if it comes in the form of an email or social media message. Fraudsters may pose as legitimate investment advisors or insiders and offer tips on assets that are "about to take off."

Huge price swings - If you see a stock that is experiencing huge price swings with no apparent reason, this could be a sign of a pump and dump scheme. 

Lack of information - If you can't find any information about a company or its products, this could be a red flag. Pump and dump fraudsters often create shell companies or use obscure assets to make it difficult for investors to do their due diligence.

Misleading information - Be skeptical of any information that seems too good to be true. Pump and dump fraudsters often make false or misleading statements about a company's financial health, products, or industry.

How to Avoid Pump and Dump Schemes?

Here are some tips for avoiding pump and dump investment scams:

Do your research - Before investing in any stock, crypto or other asset, do your research. Read the company's annual report, financial statements, and news articles about the company.

Be skeptical - If an investment opportunity seems too good to be true, it probably is. Be skeptical of high-pressure sales tactics, unsolicited investment advice, and promises of guaranteed returns.

Avoid small-cap cryptocurrencies or assets - Investors should be wary of investing in small-cap cryptocurrencies or assets, particularly those with low liquidity.

Use stop-loss orders - Investors can use stop-loss orders to automatically sell their holdings if the price of the cryptocurrency or assets falls below a certain level.

Diversify your portfolio - Don't put all your eggs in one basket. Diversify your portfolio by investing in a variety of stocks, bonds, cryptos and other assets. This will help minimize your risk and protect you from losses.

Use a reputable trading platform - When buying or selling assets, use a reputable trading platform only. Check the platform's credentials and make sure they are trusted, widely used and legally registered.

Is Pump and Dump illegal?

Yes, the pump and dump scheme is illegal as it involves the intentional manipulation of the price of assets through false or misleading information. This scheme is a violation of securities laws in many jurisdictions and can result in serious legal consequences for those who participate in it. 

In addition to legal repercussions, individuals who fall victim to the pump and dump scheme can suffer significant financial losses. It is essential for investors to be aware of the risks associated with this fraudulent activity and to exercise caution when investing in cryptocurrencies.

The Bottom Line

In conclusion, the pump and dump scheme is a fraudulent trading strategy that manipulates the price of financial assets by artificially inflating its value resulting in illegal profits for a select group of individuals. 

It is important for investors to be cautious, while making investment related decisions in the stocks, cryptos and digital assets. Since these assets can be volatile and unpredictable.

Disclaimer

The information provided on this website does not constitute investment advice, financial advice, trading advice, or any other sort of advice and you should not treat any of the website's content as such.

Token Metrics does not recommend that any cryptocurrency should be bought, sold, or held by you. Do conduct your own due diligence and consult your financial advisor before making any investment decisions.

Build Smarter Crypto Apps &
AI Agents in Minutes, Not Months
Real-time prices, trading signals, and on-chain insights all from one powerful API.
Grab a Free API Key
About Token Metrics
Token Metrics: AI-powered crypto research and ratings platform. We help investors make smarter decisions with unbiased Token Metrics Ratings, on-chain analytics, and editor-curated “Top 10” guides. Our platform distills thousands of data points into clear scores, trends, and alerts you can act on.
30 Employees
analysts, data scientists, and crypto engineers
Daily Briefings
concise market insights and “Top Picks”
Transparent & Compliant
Sponsored ≠ Ratings; research remains independent
Want Smarter Crypto Picks—Free?
See unbiased Token Metrics Ratings for BTC, ETH, and top alts.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
 No credit card | 1-click unsubscribe
Token Metrics Team
Token Metrics Team

Recent Posts

Research

API Explained: What 'API' Stands For & How It Works

Token Metrics Team
5

APIs power much of the software and services we use every day, but the acronym itself can seem abstract to newcomers. This guide answers the simple question "what does API stand for," explains the main types and patterns, and shows how developers, analysts, and researchers use APIs—especially in data-rich fields like crypto and AI—to access information and automate workflows.

What does API stand for and a practical definition

API stands for Application Programming Interface. In practice, an API is a set of rules and protocols that lets one software component request services or data from another. It defines how requests should be formatted, what endpoints are available, what data types are returned, and which authentication methods are required.

Think of an API as a contract between systems: the provider exposes functionality or data, and the consumer calls that functionality using an agreed syntax. This contract enables interoperability across languages, platforms, and teams without sharing internal implementation details.

Common API types and architectural styles

APIs come in several flavors depending on purpose and architecture. Understanding these helps you choose the right integration approach:

  • REST (Representational State Transfer): The most widespread style for web APIs. Uses HTTP methods (GET, POST, PUT, DELETE) and typically exchanges JSON. REST is stateless and often organized around resources.
  • GraphQL: A query language and runtime that allows clients to request precisely the data they need in a single request. Useful when clients require flexible access patterns.
  • gRPC: A high-performance RPC framework using protocol buffers. Favored for low-latency internal services.
  • WebSocket and Streaming APIs: For real-time, bidirectional data flows such as live price feeds or telemetry.
  • Library/SDK APIs: Language-specific interfaces that wrap lower-level HTTP calls into idiomatic functions.

In domains like crypto, API types often include REST endpoints for historical data, WebSocket endpoints for live market updates, and specialized endpoints for on-chain data and analytics.

How APIs are used: workflows and practical examples

APIs unlock automation and integration across many workflows. Typical examples include:

  • Data pipelines: scheduled API pulls ingested into analytics systems or data warehouses.
  • Automation: triggering events, notifications, or trades from software agents (when permitted by policy and regulation).
  • Embedding functionality: maps, payment processing, or identity services added to products without rebuilding them.
  • AI and model inputs: APIs provide training and inference data streams for models, or let models query external knowledge.

For researchers and developers in crypto and AI, APIs enable programmatic access to prices, on-chain metrics, and model outputs. Tools that combine multiple data sources through APIs can accelerate analysis while maintaining reproducibility.

Security, rate limits, and best-practice design

APIs must be designed with security and reliability in mind. Key considerations include:

  • Authentication and authorization: API keys, OAuth, and signed requests limit access and define permissions.
  • Rate limiting: Prevents abuse and ensures fair usage across clients; consumers should implement exponential backoff and caching.
  • Input validation and error handling: Clear error codes and messages make integrations robust and diagnosable.
  • Versioning: Maintain compatibility for existing users while enabling iterative improvements.

Designing or choosing APIs with clear documentation, sandbox environments, and predictable SLAs reduces integration friction and downstream maintenance effort.

Build Smarter Crypto Apps & AI Agents with Token Metrics

Token Metrics provides real-time prices, trading signals, and on-chain insights all from one powerful API. Grab a Free API Key

FAQ: Common questions about APIs

What does API stand for?

API stands for Application Programming Interface. It is a defined set of rules that enables software to communicate and exchange data or functionality with other software components.

How does an API differ from a library or SDK?

An API is a specification for interaction; a library or SDK is an implementation that exposes an API in a specific programming language. Libraries call APIs internally or provide convenience wrappers for API calls.

When should I use REST vs GraphQL?

Use REST for simple, resource-oriented endpoints and predictable cacheable interactions. Use GraphQL when clients require flexible, tailored queries and want to minimize round trips for composite data needs.

How do rate limits affect integrations?

Rate limits cap how many requests a client can make in a given period. Respecting limits with caching and backoff logic prevents service disruption and helps maintain reliable access.

Can APIs provide real-time data for AI models?

Yes. Streaming and WebSocket APIs can deliver low-latency data feeds that serve as inputs to real-time models, while REST endpoints supply bulk or historical datasets used for training and backtesting.

What tools help manage multiple API sources?

Integration platforms, API gateways, and orchestration tools manage authentication, rate limiting, retries, and transformations. For crypto and AI workflows, data aggregation services and programmatic APIs speed analysis.

How can I discover high-quality crypto APIs?

Evaluate documentation, uptime reports, data coverage, authentication methods, and community usage. Platforms that combine market, on-chain, and research signals are especially useful for analytical workflows.

Where can I learn more about API best practices?

Official style guides, API design books, and public documentation from major providers (Google, GitHub, Stripe) offer practical patterns for versioning, security, and documentation.

Disclaimer: This article is educational and informational only. It does not constitute financial, legal, or investment advice. Readers should perform independent research and consult appropriate professionals for their specific needs.

Research

Mastering the ChatGPT API: Practical Developer Guide

Token Metrics Team
5

ChatGPT API has become a foundational tool for building conversational agents, content generation pipelines, and AI-powered features across web and mobile apps. This guide walks through how the API works, common integration patterns, cost and performance considerations, prompt engineering strategies, and security and compliance checkpoints — all framed to help developers design reliable, production-ready systems.

Overview: What the ChatGPT API Provides

The ChatGPT API exposes a conversational, instruction-following model through RESTful endpoints. It accepts structured inputs (messages, system instructions, temperature, max tokens) and returns generated messages and usage metrics. Key capabilities include multi-turn context handling, role-based prompts (system, user, assistant), and streaming responses for lower perceived latency.

When evaluating the API for a project, consider three high-level dimensions: functional fit (can it produce the outputs you need?), operational constraints (latency, throughput, rate limits), and cost model (token usage and pricing). Structuring experiments around these dimensions produces clearer decisions than ad-hoc prototyping.

How the ChatGPT API Works: Architecture & Tokens

At a technical level, the API exchanges conversational messages composed of roles and content. The model's input size is measured in tokens, not characters; both prompts and generated outputs consume tokens. Developers must account for:

  • Input tokens: system+user messages sent with the request.
  • Output tokens: model-generated content returned in the response.
  • Context window: maximum tokens the model accepts per request, limiting historical context you can preserve.

Token-awareness is essential for cost control and designing concise prompts. Tools exist to estimate token counts for given strings; include these estimates in batching and truncation logic to prevent failed requests due to exceeding the context window.

Integration Patterns and Use Cases

Common patterns for integrating the ChatGPT API map to different functional requirements:

  1. Frontend chat widget: Short, low-latency requests per user interaction with streaming enabled for better UX.
  2. Server-side orchestration: Useful for multi-step workflows, retrieving and combining external data before calling the model.
  3. Batch generation pipelines: For large-scale content generation, precompute outputs asynchronously and store results for retrieval.
  4. Hybrid retrieval-augmented generation (RAG): Combine a knowledge store or vector DB with retrieval calls to ground responses in up-to-date data.

Select a pattern based on latency tolerance, concurrency requirements, and the need to control outputs with additional logic or verifiable sources.

Cost, Rate Limits, and Performance Considerations

Pricing for ChatGPT-style APIs typically ties to token usage and model selection. For production systems, optimize costs and performance by:

  • Choosing the right model: Use smaller models for routine tasks where quality/latency tradeoffs are acceptable.
  • Prompt engineering: Make prompts concise and directive to reduce input tokens and avoid unnecessary generation.
  • Caching and deduplication: Cache common queries and reuse cached outputs when applicable to avoid repeated cost.
  • Throttling: Implement exponential backoff and request queuing to respect rate limits and avoid cascading failures.

Measure end-to-end latency including network, model inference, and application processing. Use streaming when user-perceived latency matters; otherwise, batch requests for throughput efficiency.

Best Practices: Prompt Design, Testing, and Monitoring

Robust ChatGPT API usage blends engineering discipline with iterative evaluation:

  • Prompt templates: Maintain reusable templates with placeholders to enforce consistent style and constraints.
  • Automated tests: Create unit and integration tests that validate output shape, safety checks, and critical content invariants.
  • Safety filters and moderation: Run model outputs through moderation or rule-based filters to detect unwanted content.
  • Instrumentation: Log request/response sizes, latencies, token usage, and error rates. Aggregate metrics to detect regressions.
  • Fallback strategies: Implement graceful degradation (e.g., canned responses or reduced functionality) when API latency spikes or quota limits are reached.

Adopt iterative prompt tuning: A/B different system instructions, sampling temperatures, and max tokens while measuring relevance, correctness, and safety against representative datasets.

Build Smarter Crypto Apps & AI Agents with Token Metrics

Token Metrics provides real-time prices, trading signals, and on-chain insights all from one powerful API. Grab a Free API Key

FAQ: What is the ChatGPT API and when should I use it?

The ChatGPT API is a conversational model endpoint for generating text based on messages and instructions. Use it when you need flexible, context-aware text generation such as chatbots, summarization, or creative writing assistants.

FAQ: How do tokens impact cost and context?

Tokens measure both input and output size. Longer prompts and longer responses increase token counts, which raises cost and can hit the model's context window limit. Optimize prompts and truncate history when necessary.

FAQ: What are common strategies for handling rate limits?

Implement client-side throttling, request queuing, exponential backoff on 429 responses, and prioritize critical requests. Monitor usage patterns and adjust concurrency to avoid hitting provider limits.

FAQ: How do I design effective prompts?

Start with a clear system instruction to set tone and constraints, use examples for format guidance, keep user prompts concise, and test iteratively. Templates and guardrails reduce variability in outputs.

FAQ: What security and privacy practices should I follow?

Secure API keys (do not embed in client code), encrypt data in transit and at rest, anonymize sensitive user data when possible, and review provider data usage policies. Apply access controls and rotate keys periodically.

FAQ: When should I use streaming responses?

Use streaming to improve perceived responsiveness for chat-like experiences or long outputs. Streaming reduces time-to-first-token and allows progressive rendering in UIs.

Disclaimer

This article is for informational and technical guidance only. It does not constitute legal, compliance, or investment advice. Evaluate provider terms and conduct your own testing before deploying models in production.

Research

Mastering the OpenAI API: Practical Guide

Token Metrics Team
5

The OpenAI API has become a foundation for building modern AI applications, from chat assistants to semantic search and generative agents. This post breaks down how the API works, core endpoints, implementation patterns, operational considerations, and practical tips to get reliable results while managing cost and risk.

How the OpenAI API Works

The OpenAI API exposes pre-trained and fine-tunable models through RESTful endpoints. At a high level, you send text or binary payloads and receive structured responses — completions, chat messages, embeddings, or file-based fine-tune artifacts. Communication is typically via HTTPS with JSON payloads. Authentication uses API keys scoped to your account, and responses include usage metadata to help with monitoring.

Understanding the data flow is useful: client app → API request (model, prompt, params) → model inference → API response (text, tokens, embeddings). Latency depends on model size, input length, and concurrency. Many production systems put the API behind a middleware layer to handle retries, caching, and prompt templating.

Key Features & Endpoints

The API surface typically includes several core capabilities you should know when planning architecture:

  • Chat/Completion: Generate conversational or free-form text. Use system, user, and assistant roles for structured prompts.
  • Embeddings: Convert text to dense vectors for semantic search, clustering, and retrieval-augmented generation.
  • Fine-tuning: Customize models on domain data to improve alignment with specific tasks.
  • Files & Transcriptions: Upload assets for fine-tune datasets or to transcribe audio to text.
  • Moderation & Safety Tools: Automated checks can help flag content that violates policy constraints before generation is surfaced.

Choosing the right endpoint depends on the use case: embeddings for search/indexing, chat for conversational interfaces, and fine-tuning for repetitive, domain-specific prompts where consistency matters.

Practical Implementation Tips

Design patterns and practical tweaks reduce friction in real-world systems. Here are tested approaches:

  1. Prompt engineering and templates: Extract frequently used structures into templates and parameterize variables. Keep system messages concise and deterministic.
  2. Chunking & retrieval: For long-context tasks, use embeddings + vector search to retrieve relevant snippets and feed only the most salient content into the model.
  3. Batching & caching: Batch similar requests where possible to reduce API calls. Cache embeddings and immutable outputs to lower cost and latency.
  4. Retry logic and idempotency: Implement exponential backoff for transient errors and idempotent request IDs for safe retries.
  5. Testing and evaluation: Use automated tests to validate response quality across edge cases and measure drift over time.

For development workflows, maintain separate API keys and quotas for staging and production, and log both prompts and model responses (with privacy controls) to enable debugging and iterative improvement.

Security, Cost Control, and Rate Limits

Operational concerns are often the difference between a prototype and a resilient product. Key considerations include:

  • Authentication: Store keys securely, rotate them regularly, and avoid embedding them in client-side code.
  • Rate limits & concurrency: Respect published rate limits. Use client-side queues and server-side throttling to smooth bursts and avoid 429 errors.
  • Cost monitoring: Track token usage by endpoint and user to identify high-cost flows. Use sampling and quotas to prevent runaway spend.
  • Data handling & privacy: Define retention and redaction rules for prompts and responses. Understand whether user data is used for model improvement and configure opt-out where necessary.

Instrumenting observability — latency, error rates, token counts per request — lets you correlate model choices with operational cost and end-user experience.

Build Smarter Crypto Apps & AI Agents with Token Metrics

Token Metrics provides real-time prices, trading signals, and on-chain insights all from one powerful API. Grab a Free API Key

What are common failure modes and how to mitigate them?

Common issues include prompt ambiguity, hallucinations, token truncation, and rate-limit throttling. Mitigation strategies:

  • Ambiguity: Add explicit constraints and examples in prompts.
  • Hallucination: Use retrieval-augmented generation and cite sources where possible.
  • Truncation: Monitor token counts and implement summarization or chunking for long inputs.
  • Throttling: Apply client-side backoff and request shaping to prevent bursts.

Run adversarial tests to discover brittle prompts and incorporate guardrails in your application logic.

Scaling and Architecture Patterns

For scale, separate concerns into layers: ingestion, retrieval/indexing, inference orchestration, and post-processing. Use a vector database for embeddings, a message queue for burst handling, and server-side orchestration for prompt composition and retries. Edge caching for static outputs reduces repeated calls for common queries.

Consider hybrid strategies where smaller models run locally for simple tasks and the API is used selectively for high-value or complex inferences to balance cost and latency.

FAQ: How to get started and troubleshoot

What authentication method does the OpenAI API use?

Most implementations use API keys sent in an Authorization header. Keys must be protected server-side. Rotate keys periodically and restrict scopes where supported.

Which models are best for embeddings versus chat?

Embedding-optimized models produce dense vectors for semantic tasks. Chat or completion models prioritize dialogue coherence and instruction-following. Select based on task: search and retrieval use embeddings; conversational agents use chat endpoints.

How can I reduce latency for user-facing apps?

Use caching, smaller models for simple tasks, pre-compute embeddings for common queries, and implement warm-up strategies. Also evaluate regional endpoints and keep payload sizes minimal to reduce round-trip time.

What are best practices for fine-tuning?

Curate high-quality, representative datasets. Keep prompts consistent between fine-tuning and inference. Monitor for overfitting and validate on held-out examples to ensure generalization.

How do I monitor and manage costs effectively?

Track token usage by endpoint and user journey, set per-key quotas, and sample outputs rather than logging everything. Use batching and caching to reduce repeated calls, and enforce strict guards on long or recursive prompts.

Can I use the API for production-critical systems?

Yes, with careful design. Add retries, fallbacks, safety checks, and human-in-the-loop reviews for high-stakes outcomes. Maintain SLAs that reflect model performance variability and instrument monitoring for regressions.

Disclaimer

This article is for educational purposes only. It explains technical concepts, implementation patterns, and operational considerations related to the OpenAI API. It does not provide investment, legal, or regulatory advice. Always review provider documentation and applicable policies before deploying systems.

Choose from Platinum, Gold, and Silver packages
Reach with 25–30% open rates and 0.5–1% CTR
Craft your own custom ad—from banners to tailored copy
Perfect for Crypto Exchanges, SaaS Tools, DeFi, and AI Products