ElizaOS

What it does

ElizaOS is an open-source TypeScript framework for building autonomous AI agents. You install it, define your agent’s personality via a character file, plug in the LLMLLMLarge Language Model. A neural network trained on vast amounts of text to predict the next word in a sequence. Modern LLMs (GPT, Claude, Llama, Qwen, DeepSeek) generate human-quality text and are the foundation of most modern AI products.Like an autocomplete that read every book ever written. It has no memory of individual texts but it has absorbed the patterns of language so deeply that it can generate paragraphs that sound human. The skill is statistical, not conscious.Read more → of your choice (OpenAI, Anthropic, Google, Ollama for local inferenceInferenceRunning a trained AI model to produce an answer. Inference is what happens when you type a prompt into ChatGPT and get a response. The model takes your input, computes a best guess, and returns it.Like asking an expert for their opinion. The training was the decades they spent becoming an expert. The inference is the 30 seconds it takes them to answer your specific question.Read more →), connect it to platforms (Discord, Telegram, X, Farcaster), and deploy. The framework handles memory, state management, planning, and multi-agent coordination.

The architecture is modular. A plugin system with 90-plus official packages covers everything from blockchain integrations (Solana, EVM chains) to social media bots to RAG-powered knowledge bases. Worlds and Rooms provide multi-agent orchestration, where multiple agents can coordinate within shared workspaces. Strategic Action Chaining enables LLM-driven tool calls with parallel execution and conditional branching. Deployment options range from local installation to Docker containers to Eliza Cloud (a hosted offering from the team).

The project launched in October 2024 as “ai16z” on Solana, a provocative play on Andreessen Horowitz’s name. The concept: an AI-managed DAO where an agent called “Marc AIndreessen” would autonomously run a venture fund. The initial token saleICOInitial Coin Offering. A token sale where a project sells tokens directly to the public, usually before any product exists. ICOs dominated 2017-2018 funding and are now mostly replaced by airdrops, IDOs, or fair launches.Like a company selling shares to the public before going public, except with no SEC oversight, no audited financials, and often no product at all. The 2017 ICO boom showed why those guardrails exist in traditional finance.Read more → on DAOs.fun raised 420.69 SOL. In January 2025, Andreessen Horowitz’s Chris Dixon requested a name change due to brand confusion, and the project rebranded to ElizaOS.

In November 2025, Eliza Labs executed a tokenTokenA digital unit of value or access rights tracked on a blockchain. Tokens can represent ownership in a project, a right to use a service, a share of future revenue, or simply a tradable asset with no underlying claim.Like a physical poker chip a casino issues. The chip itself has no value. What makes it worth something is what it lets you do at the casino, what the casino has promised, and how much other people will pay you for it.Read more → migration from AI16Z to ELIZAOS at a 1:6 swap ratio, expanding total supply by 40% from 6.6 billion to 11 billion tokens. The expansion created new allocations for SAFT investors, the team, a foundation, and ecosystem funds.

The project was created by Shaw Walters, a San Francisco-based developer who previously built the Webaverse game platform (which collapsed after a treasury security breach) and MagicML, an open-source AI agent initiative. Team size at Eliza Labs is approximately 20. An arXiv paper (2501.06781) with 14 co-authors describes the framework architecture. Notable partnerships include Stanford University’s FDCI (Eliza Labs reports a grant exceeding $250,000 for AI agent research) and Chainlink (CCIP for cross-chain token bridging).

Value proposition

The pitch is the framework, not the token. ElizaOS is genuinely the most popular open-source AI agent framework in web3. 17,662 GitHub stars. 5,439 forks. 583 contributors. 90-plus plugins. MIT licensed. These are not inflated numbers. The developer adoption is real and verifiable.

Three minutes from installation to your first running agent. That is the claimed onboarding time, and it is roughly accurate. The TypeScript foundation makes the framework accessible to the largest pool of web developers. Model-agnostic design means no vendor lock-in; swap between OpenAI, Anthropic, Google, or run locally with Ollama. This flexibility is a genuine differentiator.

auto.fun, launched in April 2025, extends the platform to non-developers. Create and deploy AI agents via a no-code interface, with token launches via bonding curves on Raydium. This is the consumer-facing product that expands the addressable market beyond developers.

The counter-narrative is the token. The original AI16Z token reached an all-time high in January 2025. Adjusted for the 1:6 swap ratio, the effective ATHATHAll-Time High. The highest price a token has ever reached. ATH is usually quoted as a reference point for how far the current price has fallen (or risen) since the peak.Like the record lap time on a racetrack. It tells you what the car has been capable of at its absolute best, not what it will do today. Whether that record gets broken again depends on conditions that may or may not come back.Read more → per ELIZAOS is roughly $0.408. The token is down 99.7% from that level. The “first AI VCVCVenture Capital. Private investors who fund projects at an early stage in exchange for equity or token allocations. VC rounds are typically pre-launch, at steep discounts to any future public price, with multi-year vesting.Like angel investors in a startup who buy shares before the company goes public. They take more risk because the company might fail, so they get a better price. Once the company IPOs they can sell, and the public market pays whatever price it thinks is fair.Read more → fund” thesis, where an AI agent autonomously manages investments, hasn’t generated demonstrable returns. The DAO treasury, which once reportedly held around $10 million, has no publicly verified current value.

The token utility is thin. Governance voting is not yet live. There’s no stakingStakingLocking up a cryptocurrency to help secure a blockchain network, usually in exchange for rewards. The locked tokens act as a security deposit that can be taken away if the staker misbehaves.Like putting down a large rental deposit for an apartment. You get the money back if you behave, you earn interest while it's locked, and the landlord takes it if you trash the place.Read more → mechanism. There are no burns. No fee accrual from the framework itself. The token is effectively a speculative asset attached to a good open-source project. The potential Layer 1L1Layer 1. A base blockchain that runs its own consensus mechanism, executes transactions, and settles its own state. Bitcoin, Ethereum, NEAR, and Solana are all L1s. Anything built on top of an L1 is technically a Layer 2 or higher.Like the foundation of a building. Nothing else can exist on top until the foundation is solid. Different L1s make different tradeoffs for what kind of building they can support.Read more → blockchain launch (discussed with Saga since early 2025) would add token utility, but no concrete implementation exists. This is roadmap, not product.

Tokenomics

ELIZAOS is a multi-chain token: SPL on Solana (primary), ERC-20 on Ethereum and Base, BEP-20 on BNB Chain, bridged via Chainlink CCIP. Maximum supply: 11 billion. Total supply: 9.38 billion. Circulating supplyCirculating SupplyThe number of tokens currently in circulation and tradeable on the open market. Differs from total supply (which includes locked or unvested tokens) and max supply (the upper limit, if there is one).Like the number of cars on the road today versus the number ever produced. Some are in showrooms, some in junkyards, some still at the factory. Only the ones on the road count toward what people are actually driving.Read more →: 7.48 billion (68%) as of March 2026.

Distribution (post-migration):

• Community (original holders): 60% (1:6 swap from AI16Z, immediate unlock)
• SAFT Investors: 15% (12-month cliffCliffA waiting period at the start of a token vesting schedule during which no tokens unlock at all. After the cliff ends, tokens begin releasing according to the vesting schedule.Like a probationary period at a new job. You don't get your stock options on day one. You wait 12 months to prove you'll stick around, then everything starts unlocking normally.Read more →, then linear vestingVestingA schedule that locks up tokens allocated to insiders, investors, and team members, releasing them gradually over months or years. Vesting prevents insiders from dumping on public buyers immediately after launch.Like a new employee's stock options at a startup. You don't get all the shares on day one. They unlock over four years so you stick around and do the work rather than cashing out and leaving.Read more →; specific investors undisclosed)
• Team and Contributors: 10% (12-month cliff from October 2024, plus 24-month linear vesting via Streamflow smart contracts)
• LiquidityLiquidityHow easily a token can be bought or sold without moving the price. High liquidity means you can enter or exit large positions quickly at the quoted price. Low liquidity means even small trades can swing the market.Like the difference between selling a house and selling a share of Apple stock. The house might be worth more on paper, but finding a buyer at that price takes weeks. The Apple share converts to cash in one click.Read more → and Exchange Listings: 5.5% (607 million tokens, programmatic release)
• Foundation: 4.5% (24-month linear vesting)
• Ecosystem: 2.5% (3-month cliff, 9-month linear vesting)
• Protocol-Owned Liquidity: 2.5% (275 million tokens, programmatic deployment)

The distribution story has two chapters. Chapter one: a fair launchFair LaunchA token launch where everyone has the same access from day one. No private sale, no insider allocation, no VC discount. Tokens are distributed by mining, staking, or open public sale at a single price.Like a 100m sprint where everyone starts behind the same line at the same time. Some runners are faster, but nobody gets to start 10 metres ahead because they paid extra. The race is decided by the run, not by who bought the best position.Read more → on DAOs.fun in October 2024, no pre-mine, no VC allocation, community-first. Chapter two: a token migration in November 2025 that expanded supply by 40% and created new allocations for SAFT investors (15%), the team (10%), a foundation (4.5%), and ecosystem funds (2.5%). These allocations did not exist in the original fair launch. Original holders were diluted to fund them.

On the positive side, vesting is enforced on-chain via Streamflow smart contracts, with no early unlock discretion. This is better than handshake vesting. The SAFT cliff expires approximately October 2025, with the team cliff also passing; 24 months of linear selling pressure follows.

The post-swap ATH was on 12 November 2025, shortly after the migration completed. The effective ATH (adjusting the pre-swap peak for the 1:6 ratio) is approximately $0.408, and the token is down 99.7% from that. Listed on Binance (Alpha 2.0), Bybit, MEXC, KuCoin, Gate.io, Bitget, and Kraken. Liquidity is thin, with only around $86,000 within 2% of spot price. Micro-cap territory with high daily volume-to-market-cap ratio driven by speculation.

How to participate

Build agents. The primary participation mode. Install ElizaOS via npm or bun, define a character file, add plugins, and deploy. TypeScript or JavaScript proficiency required. Documentation at docs.elizaos.ai is thorough. Supports local development and Docker containerisation. No capital required beyond LLM APIAPIApplication Programming Interface. A structured way for one piece of software to talk to another. In DeAI, APIs let applications request inference from a model without running the model themselves.Like a waiter in a restaurant. You don't walk into the kitchen and cook your own meal. You tell the waiter what you want, they tell the kitchen, the kitchen cooks it, and the waiter brings it back. The API is the waiter.Read more → costs.

Contribute to the framework. 583 contributors and counting. Plugin development is the highest-impact contribution path. MIT licence means unrestricted use. GitHub Discussions are active on the main repo.

Use auto.fun. The no-code path. Create and deploy AI agents for social media management, DeFiDeFiDecentralised Finance. Financial services like lending, trading, and yield farming built on smart contracts instead of traditional banks or brokerages. DeFi protocols are usually permissionless and global.Like a vending machine that can give you a loan, swap your currencies, or invest your savings. Nobody is behind the counter, the rules are written into the machine itself, and anyone with money in the right format can use it.Read more → automation, and other tasks. Launch agent tokens via bonding curves on Solana. Available since April 2025, still early-stage.

Governance. Hold ELIZAOS tokens to participate in DAO proposals. Currently, governance operates through informal Discord signalling alongside the AI agent “Marc AIndreessen,” which retains final investment decision authority. A formal on-chain voting module is under development but not yet live.

Honest assessment

What works

The framework is genuinely excellent by the standards of web3 AI projects. 17,662 GitHub stars, 583 contributors, and 5,439 forks represent real developer adoption, not inflated metrics. Its plugin library (90-plus official packages) provides practical utility across LLM providers, social platforms, and blockchains. MIT licensing and TypeScript foundations make the framework maximally accessible.

The Stanford FDCI research partnership is a credible signal. According to Eliza Labs, the grant exceeds $250,000 for a three-phase research programme covering agent trust, multi-agent economics, and agent governance. This is the kind of academic collaboration that most crypto projects claim but few execute.

Credit where it’s due: the fair launch origin (DAOs.fun, 420.69 SOL raise) matters. No pre-mine, no VC allocation at inception. The community-first launch was genuine, even if the subsequent migration complicated the narrative.

Open source credentials are among the strongest in DeAIDeAIDecentralised AI. An umbrella term for blockchain-based projects that build AI infrastructure (compute, data, inference, models, agents) without a single central provider controlling the system.Like the difference between streaming a movie from Netflix and sharing it via BitTorrent. Netflix is fast and polished but one company controls what you can watch and what you pay. BitTorrent is messier but no single operator can shut you out.Read more →: 60 public repositories, MIT licence, an arXiv paper, and 17,545 commits. The HuggingFace presence is minimal (zero models, two datasets), but the GitHub activity is substantial and genuine.

What does not work

The token economy is broken. A micro-cap valuation for the “most popular AI agent framework in web3” suggests the market has concluded that being a good open-source framework doesn’t translate into token value. And the market is probably right. There’s no staking, no fee accrual, no burns, and no functioning governance vote. The token has no clear value capture mechanism.

The 40% supply expansion during migration is the central betrayal of the fair launch thesis. Original holders were diluted without formal on-chain governanceDAODecentralised Autonomous Organisation. A way to coordinate decisions and manage a treasury using token-weighted voting instead of a traditional company structure. Token holders propose and vote on changes directly.Like a shareholder-run company where every shareholder can vote on every decision, the votes are public, and the company can't do anything the shareholders don't approve. The coordination is messier than a normal company but nobody has unilateral control.Read more → approval to fund team, SAFT investor, foundation, and ecosystem allocations. Framing this as “75% community” allocation (by including ecosystem and foundation) is marketing maths. The effective insider allocation post-migration is 32% (SAFT 15% + team 10% + foundation 4.5% + ecosystem 2.5%).

Governance doesn’t exist in practice. The AI agent “Marc AIndreessen” retains final decision authority over treasury investments. Token holders can suggest but not vote. Major decisions, including the rebranding, ticker change, token migration, SAFT deals, and supply expansion, were all made unilaterally by Eliza Labs. For a project that launched as a DAO, this is a significant gap between narrative and reality.

Princeton researchers demonstrated that attackers can inject false context into an agent’s persistent memory via social media interactions, manipulating trading decisions through Sybil attacks. Eliza Labs’ response: “no direct updates were made as a result.” For agents potentially handling financial transactions, an acknowledged but unpatched vulnerability is alarming.

The risk

Shaw Walters’ prior project Webaverse collapsed after a treasury security breach. This doesn’t disqualify him, but the pattern of project launches followed by financial complications is worth noting. High dependence on a single founder creates concentration risk.

Competition is challenging from both directions. Within web3, Virtuals Protocol has a more complete tokenisation modelModelA trained neural network that takes inputs (text, images, audio) and produces outputs (more text, classifications, generated content). In DeAI the model is the thing that actually does the work.Like a very experienced apprentice who has spent years watching thousands of masters make furniture. They can't explain how they know when a joint is right, but they can make a chair that looks and functions like a Chippendale. The training is invisible. The output is what matters.Read more → with real revenue, and Olas offers stronger decentralisation with functioning on-chain governance. From the mainstream AI world, LangChain (118,000-plus GitHub stars) and CrewAI are adding web3 integrations without token overhead. ElizaOS sits between these worlds without clearly dominating either.

Regulatory risk is real. AI agents executing autonomous financial transactions create novel exposure across multiple jurisdictions. The “AI VC fund” framing raises obvious securities questions.

Thin liquidity (around $86,000 within 2% of spot) creates execution risk for any meaningful position. The token is effectively a micro-cap with correspondingly high volatility and daily volume-to-market-cap ratios driven by speculation rather than organic demand.

My position

I don’t hold ELIZAOS. The framework is genuinely good, and I respect the developer adoption metrics. But good open-source software doesn’t require a token, and ElizaOS hasn’t demonstrated why this one needs one. I would reconsider if the governance voting module ships, if the Layer 1 blockchain materialises with clear token utility, and if the memory injection vulnerability is addressed. For now, the framework is worth using; the token is not worth holding.

Freedom Score: 52/100

ElizaOS scores 52/100 (D grade). Full methodology at Freedom Score Methodology.

Infrastructure decentralisation (6/20): ElizaOS is a framework that runs on users’ own infrastructure, which is philosophically decentralised. Anyone can run agents anywhere. However, the project has no decentralised infrastructure layer: no permissionless node network, no decentralised compute marketplace, no on-chain execution environment. The DAO treasury is managed by a single AI agent with significant centralisation in Eliza Labs. The auto.fun launchpad appears centrally operated. A planned L1 blockchain via Saga would improve this score but does not yet exist.

Governance decentralisation (5/20): Governance is informal and heavily centralised around Shaw Walters and Eliza Labs. The DAO voting module has been promised but not launched. The AI agent “Marc AIndreessen” maintains final decision authority over treasury investments. The 40% token supply expansion was decided by the team without formal on-chain governance. Ticker and branding decisions were made unilaterally. Token holders can suggest investments but have no binding vote mechanism.

Token distribution fairness (7/15): The original AI16Z token launched via DAOs.fun community sale, a relatively fair launch mechanism with no pre-mine or VC allocation. However, the migration to ELIZAOS expanded supply by 40% and created new allocations: 15% SAFT investors, 10% team, 4.5% foundation, 2.5% ecosystem. The community retained 60% at the 1:6 ratio. Vesting is enforced on-chain via Streamflow, which is positive. The project started with fair launch principles but the migration diluted original holders to fund insider allocations.

Censorship resistance (10/15): The framework is fully open-source (MIT) and can be self-hosted, meaning anyone can deploy agents without permission. Code is on GitHub with 5,400-plus forks, making it practically uncensorable at the framework layer. Ollama integration supports local model execution. However, agents typically depend on centralised LLM APIs (OpenAI, Anthropic, Google) which can restrict usage. The auto.fun launchpad appears centrally operated and could theoretically censor agent launches.

Data sovereignty (10/15): Self-hosted ElizaOS agents store all data locally: memory, state, conversation history, and character files are under the user’s control. The framework does not require data to be sent to Eliza Labs servers. Users who run agents on their own infrastructure have full data sovereignty. However, the Princeton memory injection vulnerability demonstrated that agent memory can be corrupted through social media interactions. Agents using cloud LLM APIs send conversation data to third-party providers. The Eliza Cloud hosting option reduces data sovereignty.

Open source and transparency (14/15): Near the top of the field. The entire framework is MIT-licensed with 60 public repositories. 17,662 stars, 583 contributors, and 17,545 commits. An arXiv paper describes the architecture. Vesting wallets are publicly tracked. Streamflow contracts enforce vesting transparently. Documentation is thorough. The main gaps: token migration audit reports promised but not yet published, and DAO treasury operations could be more transparent.

Path to improvement

Three changes would materially increase ElizaOS’s score:

1. Ship the governance voting module. This is the single highest-impact improvement. Token holders spending real money on ELIZAOS should have binding governance rights. The current model, where an AI agent has final decision authority and the team makes all major decisions unilaterally, contradicts the DAO thesis. On-chain governance with community proposals, transparent voting, and binding outcomes should be the minimum standard before asking people to buy governance tokens.
2. Address the memory injection vulnerability. Princeton researchers publicly disclosed a mechanism to manipulate agent trading decisions via memory injection, and Eliza Labs responded with “no direct updates were made.” For a framework whose agents handle financial transactions, this is unacceptable. Publishing a formal security response, implementing mitigations, and commissioning a third-party audit of the core framework would demonstrate the security maturity that institutional adopters require.
3. Establish clear token utility. The ELIZAOS token currently has no value capture mechanism: no staking, no fee accrual, no burns, no governance voting. Either the L1 blockchain needs to ship with ELIZAOS as its native token, or auto.fun fees need to accrue to token holders, or some other concrete mechanism needs to create sustained demand beyond speculation. Great open-source software doesn’t need a token; if you issue one, it needs a reason to exist.

Returns Score: 27/100

ELIZAOS scores 27/100 (F grade). Full methodology at Returns Score Methodology.

Token utility (8/20): The framework is genuinely excellent: 17,600-plus GitHub stars, 583 contributors, MIT licensed, three minutes to a running agent. None of that has anything to do with the token. ELIZAOS has no staking mechanism, no fee accrual, no burns, and governance voting is not yet live. The token exists alongside the framework rather than being integrated into it. A planned L1 blockchain would change this, but plans are not products. Today, you can use the entire ElizaOS framework without ever touching the token.

Value accrual (3/20): There’s no clear mechanism by which framework adoption drives token value. The framework is open source and free to use. auto.fun generates some fees, but the connection to ELIZAOS token holders is undefined. The original “AI VC fund” thesis (where an AI agent autonomously manages investments) hasn’t produced demonstrable returns. The DAO treasury value is unverified. This is the core problem: brilliant open-source adoption with zero value capture at the token layer.

Supply dynamics (6/20): The November 2025 migration expanded total supply by 40%, from 6.6 billion to 11 billion tokens. Original holders were diluted to fund SAFT investors (15%), team (10%), foundation (4.5%), and ecosystem (2.5%). At 11 billion tokens, the per-token value is highly diluted even at modest market caps. Vesting is enforced on-chain via Streamflow, which is a positive, but the dilutive expansion itself was decided unilaterally by the team without governance vote. That sets a precedent where supply can be expanded at management’s discretion.

Revenue sustainability (2/25): There is no revenue model for the token. The framework is open source and generates no licensing fees. auto.fun is the closest thing to a revenue product, but its fee structure and relationship to the ELIZAOS token are not formalised. The “AI VC fund” has no published portfolio performance. This scores near the bottom because there is simply nothing to point to: no revenue, no fees, no concrete plan for either. The token economy runs on speculation and narrative momentum alone.

Liquidity and access (8/15): ELIZAOS trades across Solana, Ethereum, and Base, providing multi-chain access. However, daily volume has declined 99.7% from ATH levels, leaving thin order books and significant slippageSlippageThe difference between the expected price of a trade and the price you actually get when the trade executes. Slippage usually goes against the trader and gets worse with bigger trades or thinner markets.Like trying to buy 1000 bananas at the corner shop. The first few are at the marked price, but by the time you've bought them all you've moved the price up because there are no more bananas left at the original level. The shop has to restock at higher cost.Read more → risk. The token is firmly in micro-cap territory. Within 2% of spot price, only about $86,000 in liquidity sits on the book. That means any position above a few thousand dollars faces meaningful execution challenges.

Path to improvement

Three changes would materially increase ElizaOS’s returns score:

1. Ship the L1 blockchain with ELIZAOS as native gasGasThe fee paid to a blockchain to process a transaction. Gas is denominated in the chain's native token and varies with network demand. Sending a transaction without enough gas means the transaction fails and the gas is still consumed.Like the petrol that powers a car. You need to put petrol in to make the engine run. The amount of petrol you need depends on how far you're driving and how much you're carrying. If you run out, the car stops.Read more → token. This is the single change that would most directly address the token utility vacuum. If every agent deployment, every transaction, and every governance vote requires ELIZAOS, then framework adoption finally drives token demand. Without this, the token remains a speculative sidecar to a good open-source project.
2. Formalise auto.fun fee accrual to token holders. Route a percentage of auto.fun launch fees to ELIZAOS stakers or implement a buy-and-burn mechanism funded by platform revenue. This creates a concrete, auditable link between platform activity and token value that doesn’t currently exist.
3. Publish DAO treasury holdings and investment performance. The “AI VC fund” thesis was the original reason the token existed. If the fund has made investments, publish the portfolio. If it hasn’t, retire the narrative. Transparency about what the treasury holds and how it has performed would either validate or honestly retire the founding thesis.

Score change log