★ Open Source// LOCAL-FIRST OBSERVABILITY RUNTIME

You own
your data.
We provide
the runtime.

One agent. Any chain. Full control. Local-first observability for onchain and off-chain systems — no ingestion fees, no vendor lock-in, no surprises.

Get Started Free →$ cargo install loghaven

Ingestion fees

Chains supported

100%

Data ownership

loghaven — prodLIVE

$ loghaven start --profile prod

◈ LogHaven v0.3.0 starting...

  ✓ Config loaded        ~/.loghaven/config.toml

  ✓ Storage backend      r2://my-bucket/logs

  ✓ EVM adapter          mainnet, base, arbitrum

  ✓ Solana adapter       mainnet-beta

  ✓ Ingest endpoint      http://localhost:7430

● Agent running  •  Watching 4 chains  •  0 events buffered

$ ▋

Agent activeETH + BaseSolana0 errors

// Supported chains & integrations

Ethereum

Base

Arbitrum

Optimism

Solana

Stellar

AWS S3

Cloudflare R2

MinIO

Local FS

Ethereum

Base

Arbitrum

Optimism

Solana

Stellar

AWS S3

Cloudflare R2

MinIO

Local FS

Ethereum

Base

Arbitrum

Optimism

Solana

Stellar

AWS S3

Cloudflare R2

MinIO

Local FS

// THE PROBLEM

Observability tooling
broke the contract.

Modern tools charge per byte ingested and per day retained. Debug a production incident and get a $4,000 bill.

They store your production logs — and your users' data — in infrastructure you don't control.

LogHaven flips the model. The runtime is ours. The data is yours. Forever.

Existing ToolsLogHaven

Storage✗Vendor cloud✓Your storage

Ingestion cost✗Per GB✓Free

Chain support✗Single✓Multi-chain

Data ownership✗Vendor✓You

Lock-in✗High✓Zero

// WHAT IT DOES

Your agent is smart.
Your data should stay home.

Tx0x4a3f...b92d
ChainEthereum Mainnet
Status✓ Confirmed
Block21,847,302
Gas142,847
Logs3 events
Traces7 spans

Transaction-Centric

Every log, trace, and metric anchored to a transaction hash. One tx → full system context, across chains.

ETH tx

──→

Bridge

──→

Base tx

──→

Off-chain log

Cross-Chain Correlation

Link an Ethereum transaction to a Solana instruction to an off-chain API trace. Built for bridges and multi-chain apps.

observe:
  chain: solana
  when:
    tx_error_rate: "> 2%"
  then:
    enable:
      logs: debug
      traces: full
    duration: 15m

Observability as Code

Define what to watch, when to watch it, and for how long. Version-controlled. Reproducible. Zero config drift.

You: What caused the spike at 14:32?
◈  Analyzed 847 transactions across 3 chains.
   Root cause: nonce gap on Arbitrum deployer.
   Affected: 12 txs in 4-minute window.
   Similar pattern: 2024-11-03 deploy.

AI Copilot — Your Data Stays Local

Explain failures, cluster errors, compare deploys. Works with local LLMs (Ollama) or hosted models — raw logs never leave your machine by default.

// USE CASES

Built for every layer
of the stack.

Debug anything. In minutes, not hours.

Trace a failed transaction, correlate on-chain failures with off-chain logs, and ask your local LLM what happened — without sending raw logs anywhere.

◈Trace any tx across chains with a single command

◈Auto-correlates on-chain + off-chain events by hash

◈CLI-first — fits your existing terminal workflow

◈AI Copilot runs locally via Ollama or any model

loghaven — developers

$ loghaven trace 0x4a3f...b92d

◈ Fetching trace for 0x4a3f...b92d...

  Chain:          Ethereum Mainnet

  Status:         REVERTED

  Revert reason:  'InsufficientLiquidity'

  Gas used:       89,234 / 200,000

  ✓ 3 correlated off-chain logs found

  ✓ Similar failure pattern: 2024-11-03

// PHILOSOPHY

We believe observability
should belong to you.

◈

User-Owned

Your data belongs to you, not a SaaS vendor. LogHaven writes to storage you control — and only you can access.

▣

Cost-Predictable

No surprise bills from ingestion spikes. Pay for storage, not for observing your own infrastructure.

◎

Privacy-Respecting

Sensitive production logs stay on your infrastructure. Client-side encryption by default, keys never leave your machine.

⬡

Chain-Agnostic

Web3 development spans multiple chains. LogHaven treats Ethereum, Solana, and Stellar as peers — not afterthoughts.

◉

Composable

Build custom workflows, not rigid dashboards. Connect it to your existing stack — Grafana, Prometheus, Slack, anything.

✦

Open by Default

The agent is open source. Every byte that flows through it is inspectable. No black boxes in your observability stack.

// TO BE CLEAR

LogHaven is NOT any of these things.

✗

A managed SaaS platform

We don't host your data. We never will.

✗

A replacement for APM tools

We complement Datadog, Grafana, and others. We don't replace them.

✗

Blockchain-only

We support any chain via adapters. Off-chain logs are first-class citizens.

✗

A closed ecosystem

Bring your own storage, your own tools, your own workflows.

// ARCHITECTURE

One runtime.
Wired to everything.

A long-running daemon handles all ingestion. Your CLI, TUI, and dashboard talk to it via IPC — lightweight, crash-tolerant, and always on.

USER INTERFACES

CLITUIDashboardREST API

IPC — Unix Socket / Named Pipe

DAEMON RUNTIME — CORE

Ingestion Pipeline

Normalize · Buffer · Validate

Transaction Indexer

Hash · Correlate · Store

Event Processing Engine

Filter · Enrich · Route · Alert

Chain Adapters

Ethereum

Solana

Stellar

Base · Arbitrum · ...

Storage Engine

Local Filesystem

AWS S3

Cloudflare R2

MinIO

Daemon-based runtimeIPC via Unix socket / named pipePlugin chain adaptersPluggable storage backends

// GET STARTED

Install on any platform.
Running in under a minute.

LogHaven runs natively on Windows, macOS, and Linux. Install via your preferred method — package manager, shell script, or build from source with Cargo.

terminal — linux

# Download and install via shell script

$curl -sSL https://install.loghaven.dev | sh

# Or with wget

$wget -qO- https://install.loghaven.dev | sh

# Verify installation

$loghaven --version

# Initialize your config

$loghaven init

# Start the daemon

$loghaven run

~/.config/loghaven/config.toml

[meta]
version = 1
 
[agent]
name    = "loghaven-agent"
log_level = "info"
 
[storage]
backend = "local"
primary = "local"
 
[storage.local]
path       = "~/.loghaven/data"
max_size_gb = 10
 
[chains]
enabled = ["ethereum", "solana"]
 
[chains.ethereum]
rpc_url = "https://eth-mainnet.g.alchemy.com/v2/YOUR-KEY"
 
[chains.solana]
rpc_url = "https://api.mainnet-beta.solana.com"

◈

Use loghaven init --profile prod to create named profiles for different environments. Config files live in ~/.config/loghaven/ on Unix and %APPDATA%\loghaven\ on Windows.

// TRY IT OUT

See it in action.

Pick a scenario and watch the agent work.

Select scenario

loghaven — interactive demo

Ready to run this on your own infrastructure?

Install LogHaven →View on GitHub

// BRING YOUR OWN STORAGE

Your bucket.
Your rules.

LogHaven writes directly to storage you control. S3, R2, MinIO, or local disk. No raw data ever touches our infrastructure. Add lifecycle policies to auto-delete. Rotate keys anytime.

◈Cloudflare R2

Zero egress fees. Global edge storage.

[storage]
backend = "r2"
bucket  = "my-logs"

▣AWS S3

Battle-tested. Lifecycle policies built-in.

[storage]
backend = "s3"
bucket  = "prod-logs"
region  = "us-east-1"

▦MinIO

Self-hosted. Full S3 compatibility.

[storage]
backend  = "minio"
endpoint = "http://minio:9000"
bucket   = "logs"

◉Local Filesystem

Fastest option. Zero config. Air-gapped.

[storage]
backend = "local"
path    = "~/.loghaven/data"

// OBSERVABILITY AS CODE

Define it once.
Run it anywhere.

Declare observation rules in YAML. Commit them to your repo. The agent enforces them automatically. No dashboards to configure. No SaaS to log into.

loghaven.observe.yaml

# loghaven.observe.yaml
observations:
  - name: solana-error-spike
    chain: solana
    trigger:
      metric: tx_error_rate
      threshold: "> 2%"
      window: 5m
    response:
      enable:
        logs: debug
        traces: full
      duration: 15m
      notify:
        - channel: stderr
        - webhook: "${ALERT_WEBHOOK}"

  - name: evm-gas-anomaly
    chain: evm
    trigger:
      metric: avg_gas_price
      threshold: "> 150 gwei"
    response:
      enable:
        logs: warn
      duration: 30m

Version-controlledReproducibleNo SaaS required

◈

ENCRYPTED

════ KEY: [YOU] ════

CLIENT

NEVER PLAINTEXT

// SECURITY & TRUST

Even if compromised,
we can't read your logs.

Data is encrypted before it leaves the agent. You hold the keys. We never see plaintext.

 _         ___     ____   _   _      _      __   _____  _   _ 

| |       / _ \   / ___| | | | |    / \    \ \ / / ___|| \ | |

| |      | | | | | |  _  | |_| |   / _ \    \ V /| |_  |  \| |

| |___   | |_| | | |_| | |  _  |  / ___ \    | | |  _| | |\  |

|_____|   \___/   \____| |_| |_| /_/   \_\   |_| |____||_| \_|

// LOCAL-FIRST OBSERVABILITY RUNTIME

v0.3.0-alpha · MIT License · Built in public

★ Free & Open Source Forever

Observe freely.
Own everything.

Install in under 60 seconds. No account. No credit card. No ingestion fees. The runtime runs on your machine — and your data never leaves it.

Get Started Free →

Read the Docs →View on GitHub →

MIT Licensed · Built in public · Your data never leaves your machine

You ownyour data.We providethe runtime.

Observability toolingbroke the contract.

Your agent is smart.Your data should stay home.

Transaction-Centric

Cross-Chain Correlation

Observability as Code

AI Copilot — Your Data Stays Local

Built for every layerof the stack.

Debug anything. In minutes, not hours.

We believe observabilityshould belong to you.

User-Owned

Cost-Predictable

Privacy-Respecting

Chain-Agnostic

Composable

Open by Default

LogHaven is NOT any of these things.

One runtime.Wired to everything.

Install on any platform.Running in under a minute.

See it in action.

Your bucket.Your rules.

Define it once.Run it anywhere.

Even if compromised,we can't read your logs.

Client-Side Encryption+

Short-Lived Tokens+

Zero Raw Data Retention+

Fully Auditable Pipeline+

Observe freely. Own everything.

You own
your data.
We provide
the runtime.

Observability tooling
broke the contract.

Your agent is smart.
Your data should stay home.

Built for every layer
of the stack.

We believe observability
should belong to you.

One runtime.
Wired to everything.

Install on any platform.
Running in under a minute.

Your bucket.
Your rules.

Define it once.
Run it anywhere.

Even if compromised,
we can't read your logs.

Observe freely.
Own everything.