Skip to the Quickstart paths

Quickstart

Start with a safe, useful first step.

You do not need to be a developer to inspect Vaultfire. Choose a browser check, give a read-only brief to an AI agent, or run the actual developer example on your own computer.

No path below asks for a personal private key, a funded wallet or a V3 payment.

Choose your path

People, agents, and builders start differently.

I just want to try it

Read the public release and service information in this browser. No installation, account or wallet.

Run the no-install check

I use an AI assistant

Use a tool-capable app with a local or cloud model, or give your assistant the public documents. Five read-only tools; no signing or payment.

Connect a local AI safelyGet the read-only agent brief

I want to build

Download the exact source, use the pinned tools, and run register/revoke plus a real local HTTP/SQLite receipt flow.

Follow the developer setup

I want to understand the idea

No prerequisites, installation or funding. Read one task story and the three-pot example. If a term is unfamiliar, the glossary explains it.

Read the plain-English guide

I evaluate a partner integration

Bring your network, asset, custody and recovery requirements. Reading is free; the result is a gap and acceptance review, not activation. Keep the exact source identity and unresolved gates.

Inspect partner readiness

I evaluate agent finance

No funds needed to inspect the scope. The executable journey requires the pinned local tools. Preserve its operation ID and journal to recover; never retry a public payment to make it pass.

Review permissions and gaps

Six-target local evaluation

Choose your target. Keep its boundaries.

The simple default Quickstart below remains separate. These paths use disposable local assets and a real isolated private service, not public payments or a shared cross-chain balance.

Selected target: Base. Public financial activation remains disabled.

Six chains. Separate evidence. No public V3 payments.

All six targets have source and local evaluation paths. Native Solana is not an EVM profile. Historical local results are not fresh acceptance of this source, production qualification or permission to activate.

  • BaseEVM · 8453

    Implementation
    V3 Solidity plus target-specific local finance adapter; separate asset and fee holds.
    Local evidence
    Historical N04 protocol/private-finance composition. Uses an isolated target profile and disposable assets. N05 final acceptance is separate.
    Public reads / writes
    Existing V2/read-only surfaces remain distinct. Public V3 financial writes disabled.
    Remaining gates
    Qualify execution, L1 data and historical operator fees, asset implementation and finality independently. Exact asset, custody, provider/finality, independent review and operator qualification remain required.
  • SolanaNative program

    Implementation
    Native Rust successor and helper; SOL and separate rent/fee liabilities.
    Local evidence
    Historical N04 SBF build, program-test, validator and private-finance evidence. Fresh source-bound rerun remains a separate gate.
    Public reads / writes
    Existing V2/read-only surfaces remain distinct. Public V3 financial writes disabled.
    Remaining gates
    Local genesis is generated per rehearsal. Intended mainnet genesis is 5eykt4UsFv8P8NJdTREpY1vzqKqZKvdpKuc147dw2N9d; public custody is not qualified. Native SOL does not establish SPL support. Exact asset, custody, provider/finality, independent review and operator qualification remain required.
  • ArbitrumEVM · 42161

    Implementation
    V3 Solidity plus target-specific local finance adapter; separate asset and fee holds.
    Local evidence
    Historical N04 protocol/private-finance composition. Uses an isolated target profile and disposable assets. N05 final acceptance is separate.
    Public reads / writes
    Public discovery is not V3 runtime acceptance. Public V3 financial writes disabled.
    Remaining gates
    Qualify canonical finality and parent-fee representation without double-counting. Exact asset, custody, provider/finality, independent review and operator qualification remain required.
  • PolygonEVM · 137

    Implementation
    V3 Solidity plus target-specific local finance adapter; separate asset and fee holds.
    Local evidence
    Historical N04 protocol/private-finance composition. Uses an isolated target profile and disposable assets. N05 final acceptance is separate.
    Public reads / writes
    Public discovery is not V3 runtime acceptance. Public V3 financial writes disabled.
    Remaining gates
    Distinguish the selected execution-finality policy, checkpoints and bridge claims. No bridge credit is admitted. Exact asset, custody, provider/finality, independent review and operator qualification remain required. The N05 original admission/chain-effect incident remains unresolved; a retry is not closure.
  • AvalancheEVM · 43114

    Implementation
    V3 Solidity plus target-specific local finance adapter; separate asset and fee holds.
    Local evidence
    Historical N04 protocol/private-finance composition. Uses an isolated target profile and disposable assets. N05 final acceptance is separate.
    Public reads / writes
    Public discovery is not V3 runtime acceptance. Public V3 financial writes disabled.
    Remaining gates
    Qualify the actual upgrade, charged-gas receipt semantics, asset behavior and wallet debits. Exact asset, custody, provider/finality, independent review and operator qualification remain required.
  • MonadEVM · 143

    Implementation
    V3 Solidity plus target-specific local finance adapter; separate asset and fee holds.
    Local evidence
    Historical N04 protocol/private-finance composition. Uses the separate pinned Monad engine, not a relabeled Anvil. N05 final acceptance is separate.
    Public reads / writes
    Public discovery is not V3 runtime acceptance. Public V3 financial writes disabled.
    Remaining gates
    Requires the separately pinned anvil-monad-v1.8.7 engine. A baseline Anvil process with chain ID 143 is not Monad acceptance. Exact asset, custody, provider/finality, independent review and operator qualification remain required.

Source and evidence identity: the current source selector binds the immutable protocol archive; the publication identity binds the website separately. Manifest e3c24f8812f4f5a2530e4aa2bcc8ff01c29c3d163a70306c9a266f69fca608f6. Review module-level implementation and evidence before interpreting a chain profile as full feature parity.

Base: chain ID 8453, disposable MockERC20

The local token has 18 decimals; accounting uses integer base units only. This does not qualify USDC, another public token, public fees or public finality.

Prerequisites and installation

Use the verified current source and its pinned Linux environment: Node 20.20.1, npm 10.8.2, Foundry 1.3.1, Solc 0.8.24, Python 3.14.3, SQLite CLI 3.46.1, bubblewrap and util-linux. Real Unix sockets and network, mount and PID namespaces are required; do not replace isolation with a mock.

Minimal SDK client and connected journey

The first client command genuinely imports the built SDK and prints the selected read-only profile. It performs offline inspection, not RPC, estimation, signing, submission or settlement. The following separate workflow exercises supported local finances.

Base local client and financial journey
npm --prefix repos/vaultfire-sdk ci
npm --prefix full-v3/sdk ci
npm --prefix full-v3/sdk run build
node full-v3/sdk/examples/target-client.mjs base
export VAULTFIRE_PYTHON="$(command -v python3.14)"
"$VAULTFIRE_PYTHON" scripts/run_protocol_composition.py --chain-id 8453 --output-dir "/tmp/vaultfire-base-$(date +%s)"

Expected output and timings

Client output must identify vaultfire.target-client.v1, the selected target and publicPayments: false. The composition must exit zero with decision PASS in acceptance.json, both protocol and finance child reports, unchanged execution inputs and a shared-bond-finance report.

The November 4 prepared local run recorded 97.803 seconds summed across its actual composition commands. These are scoped engineering measurements, not a guarantee or fresh acceptance of a later archive. First-time tool downloads and installation were not independently timed here. The separate prepared default Quickstart measured 39.324 seconds; full six-target acceptance is not a five-minute task.

Failure diagnosis and recovery

Stop at the first failed parent and retain its complete output. Check exact tool versions, executable Python/SQLite, free capacity, namespace permissions, localhost port availability and asset runtime, chain profile and engine identity. Never loosen a guard, reuse a failed output directory or claim a child pass completes its parent.

Preserve operation IDs, nonce/wire, durable receipts and unresolved holds. Inspect status before retrying; an interruption must not authorize a second payment. Keep private wallets, signing keys, ledgers and checkpoints local and out of public artifacts. An offline restore must not automatically resume payments.

Still required for this public target

Qualify execution, L1 data and historical operator fees, asset implementation and finality independently. Qualify exact assets, signer, providers, fees, finality, staffed operations, independent review and any required genuine proof before requesting explicit activation approval.

No installation needed

Read the current public information.

This button reads three public pages: the release record, the agent discovery document, and service health. It does not connect a wallet, sign, pay, or submit a V3 transaction.

If nothing loads, keep the error message and check your connection. Open the source record directly, inspect discovery, or read reported service health.

Local AI path

Your model can run on your computer.

A local model does the AI thinking on your own device. A tool-capable application connects it to Vaultfire. You do not need a cloud AI provider, a wallet or a payment key for the five public read-only MCP tools.

  1. Load a local model: use LM Studio with a model that supports tool use. Model downloads and hardware needs vary; choose a local model rather than a cloud-backed one.
  2. Add the connection: in LM Studio, open Program, then Install > Edit mcp.json. Merge the server entry below into your existing mcpServers object; keep any unrelated entries.
  3. Try a safe question: ask your model to use vaultfire_list_operations, then vaultfire_get_operation for /api/x402/trust/contracts, and explain what is available without paying or signing.
  4. Look for a real tool result: discovery should describe 76 operations, including two free ones. Reading the description of a paid operation does not execute or pay for it. A model answer without a visible tool call is not proof that a request ran.
LM Studio read-only MCP configuration
{
  "mcpServers": {
    "vaultfire-read-only": {
      "url": "https://theloopbreaker.com/api/mcp"
    }
  }
}

Local inference is not the same as offline operation. This connection sends tool requests to the public Vaultfire server, not to a cloud model provider. Tool arguments and normal request metadata leave your device. Never put private keys, wallet seeds or confidential conversation text in those arguments.

The configuration grants no signing, payment or filesystem tools. Standard MCP transport tests do not certify every local model or app version. See LM Studio's setup instructions and verify the result on your own device.

Developer path

Build from exact source, not a guessed example.

This path uses a terminal, which is the application where you type commands for your computer to run. If that is unfamiliar, you can use the browser check above first, or share these steps with a developer. Do not type commands into a wallet or a chat window and assume they ran.

  1. Prepare: use the included Linux development container, or install the exact versions listed below. Installation can take longer than the example itself. Other host setups are not implied to be tested.
  2. Download and verify: obtain the selected source and its checksums. A matching checksum means the bytes match the published package, not that the software is independently audited.
  3. Run locally: install both packages and execute the existing Quickstart. It signs only with disposable public test keys on a local test chain, never your personal wallet or a public network.
  4. Check success: the final report must have top-level "status": "passed", show a consent record being registered and revoked, and show the local receipt reaching finalized. Keep the report and source identity.
  5. Build your integration: follow the versioned SDK and API operation contracts. Check which operations are supported before adding authorization, retries or status polling. Do not treat a schema or simulated catalog as live execution.

Extended local evaluation

Run the connected treasury, payment and escrow journey.

This is a separate maintained workflow, not a replacement for the Quickstart below. Start from the verified source with both SDK lockfiles installed. Use pinned Node 20.20.1, Foundry 1.3.1, Solc 0.8.24, Python 3.14.3 and SQLite CLI 3.46.1 on Linux, with real bubblewrap and util-linux for process isolation. Python’s embedded SQLite and the isolation-tool versions are reported independently.

Maintained local protocol and financial composition
npm --prefix repos/vaultfire-sdk ci
npm --prefix full-v3/sdk ci
# Install the exact repository toolchain first; no personal wallet or keys.
export VAULTFIRE_PYTHON="$(command -v python3.14)"
"$VAULTFIRE_PYTHON" scripts/run_protocol_composition.py --output-dir "/tmp/vaultfire-composition-$(date +%s)"

The runner builds both SDKs and exact Solidity artifacts, then runs the maintained protocol and financial workflows. Within the finance run, invoices, treasury, escrow, bonds, rewards, optional native cover and recovery share one actual local chain and private service. Look for top-level decision: PASS in acceptance.json, exact before/after input bindings, both child reports and shared-bond-finance-report.json. Delivery/scores are synthetic; wallets and freely minted tokens are disposable. No public funds or paid provider are used.

Your costs are local compute, storage and prerequisite downloads. Installation and compilation can take substantially longer than execution. Preserve a failed output directory and its logs; choose a new directory for a complete rerun. For interrupted real submissions, inspect durable status and the receipt first rather than rebroadcasting.

Developer Quickstart · local only

Run an example, then check its receipt.

Register consent, revoke it, and inspect a local receipt. This example runs on your computer with disposable local services. It is not a Base mainnet or Solana deployment, and it needs no funded wallet, private key or payment.

Not ready to install development tools? Start with either path below. A terminal is the window where these commands run on your own computer.

  1. Get the current source. Use the current download and follow its checksum and extraction checks before running it. Do not substitute an older bundle if retrieval fails.
  2. Install the pinned environment. Use the partner devcontainer included at full-v3/partner/.devcontainer, or install the exact versions below. Dependency installation and native tool installation are separate setup costs, not part of a prepared quickstart timing.
  3. Run only locally. From the verified extracted root, install the SDK dependencies as well as the sandbox, then run the existing quickstart. No wallet, private key, seed phrase, payment header or public RPC is required or accepted.
local-only quickstart commands
# From the authenticated extracted source root, with pinned tools installed:
npm --prefix full-v3/sdk ci --ignore-scripts
npm --prefix full-v3/partner ci --ignore-scripts
npm --prefix full-v3/partner run quickstart

What success looks like: the final report has top-level "status": "passed", confirms a local consent record was registered and revoked, and shows the local receipt reaching finalized. Keep the output. A pass proves this local example ran; it does not activate V3 or qualify partner funds.

If it fails: keep the error and stop at the failing step. Missing tools require the exact pinned environment, not looser checks. A busy local port must not be replaced with a public network. Package-download failures require restoring the documented download path, not sharing credentials. Recovery details are below.

Alternative retrieval: pinned Twigpine clone

Authenticate the selected protocol commit after cloning. If the transport is unavailable, use the current HTTP archive above; never switch silently to a historical release.

export VAULTFIRE_TWIGPINE_TOOLS="$(mktemp -d)"
npm install --prefix "$VAULTFIRE_TWIGPINE_TOOLS" --no-save --ignore-scripts=false @gitlawb/gl@0.7.1
export PATH="$VAULTFIRE_TWIGPINE_TOOLS/node_modules/@gitlawb/gl/bin:$PATH"
test "$(git-remote-gitlawb --version)" = "git-remote-gitlawb 0.7.1"
test "$(uname -s)-$(uname -m)" != "Linux-x86_64" || test "$(sha256sum "$(readlink -f "$(command -v git-remote-gitlawb)")" | cut -d' ' -f1)" = "af5da2da412dfe24fff57b1d8a0834da1fa5d971897423a19ded454a0f732e74"
export GITLAWB_NODE=https://node.gitlawb.com
git clone --recurse-submodules gitlawb://did:key:z6MkryiNsYdFEMHv95wxzSQ1vtFXPyVQQrxXdPw3tE5HpfxV/vaultfire
cd vaultfire
git checkout --detach cd5956d6fec7b7a9f42bbb10468ff393ce040b11
test "$(git rev-parse HEAD)" = "cd5956d6fec7b7a9f42bbb10468ff393ce040b11"
npm --prefix full-v3/sdk ci --ignore-scripts
npm --prefix full-v3/partner ci --ignore-scripts
npm --prefix full-v3/partner run quickstart

Pinned environment: Node 20.20.1, npm 10.8.2, SQLite 3.46.1, Foundry 1.3.1, and project Solc 0.8.24. This is local software verification only. It does not deploy V3 to Base or Solana, enable payments, verify a real ML-DSA proof, or authorize partner funds.

Verify the download and inspect implementation details

The current source selector binds the selected archive to its checksum and external sidecar inventory. Verify every extracted file with the supplied Git-free verifier. An archive takes its identity from its own authenticated external sidecar, not an embedded older website projection. Protocol source identity and website publication identity are separate.

The local example builds the source SDK, compiles ConsentRegistryV3, runs register/revoke against disposable loopback Anvil, and finalizes a receipt through loopback HTTP and SQLite.

Archive SHA-256: 8ff8ad69deb666d9fba1556bece4257b74cdc584b2dedacc9381705ce348a1c6. The current HTTP archive is the first path; clone transport is optional. Historical source is never an automatic fallback.

Expected result, recovery and full-validation limits

A passing quickstart prints its scoped command results, register/revoke reads and local receipt result. Retain the actual stdout, exit code, source identity, environment and elapsed time. Reading these instructions is not execution evidence.

  • Version/tool missing: use the pinned partner devcontainer; do not loosen version checks. The website separately requires Node 22, while the partner sandbox pins Node 20.20.1.
  • Install/network failure: keep the failure log and retry only after restoring the documented package/tool transport. Never share proxy credentials or private CA material with this site.
  • Quickstart fails or times out: keep the source unchanged and report the failing command. A retry in a warmed environment is prepared execution, not cold installation evidence.
  • Full validation: npm --prefix full-v3/partner run full-audit is separate and may require native Solana/Anchor, Rust, cryptography/proof toolchains and substantial resources. Host/simulation checks do not close native, real-proof, release or deployment gates.
Vaultfire

Morals over metrics. Privacy over surveillance. Freedom over control.

Inspect before you trust
Vaultfire Quickstart | People, agents and builders