ObsidianLog encrypts, compresses, and hash-chains your logs before archiving them to Sia's decentralized storage network, so old logs cost next to nothing to keep, and you can prove nothing was quietly changed or deleted.
Every application and server you run produces logs constantly. Tools like Datadog and Elastic are great for watching them right now, but they charge by the gigabyte and by the month, so once logs aren't being actively watched anymore, most teams just delete them, even though audits, incident investigations, and compliance rules often call for keeping them much longer.
ObsidianLog gives your old logs a permanent, cheap home — encrypted before they ever leave your machine, so only you can read them, and cryptographically chained together, so you can prove later that nothing was quietly changed or deleted. Your day-to-day monitoring stays exactly as it is. ObsidianLog is just where logs go once you're done watching them in real time.
AES-256-GCM, with a key you generate and hold yourself. It's never sent to Sia, to a storage host, or to us.
Every entry is SHA-256 hash-chained to the one before it. Run obsidianlog verify any time and know immediately if anything was altered.
A fraction of what equivalent storage costs on Datadog, Elastic, or AWS, so nothing has to get deleted just to control cost.
A Vector-compatible HTTP endpoint. Point your existing log shipper at it, no new agent, no code changes.
Hosted by default via sia.storage, so there's no server for you to babysit. Bring your own indexer if you want full control instead.
Inspect exactly what it does, self-host every part of it, or fork it. Nothing about your archive depends on us staying in business.
Built for technically capable developers and small teams comfortable with a CLI and a config file.
You probably aren't archiving logs anywhere today. This gives you real, tamper-evident archival for close to nothing.
You need to prove logs weren't altered after the fact, not just store them somewhere. That's what the hash chain is for.
You're already using Sia. ObsidianLog is a natural addition to your existing workflow.
Owning your own data instead of trusting a vendor with it isn't just a preference, it's the whole point.
A cheap, durable place to keep CI logs, release history, and telemetry around for good, not just a few weeks.
One pipeline, four commands. Here's what happens to a log line, start to finish, and exactly what each command does.
Run this once. A short wizard generates your encryption key, asks where you want data stored (locally to try it out, or on Sia for the real thing), and saves your configuration. Nothing here requires you to understand Sia, blockchains, or wallets.
Starts a small local server that listens for log batches from your existing log shipper (like Vector) and runs them through the pipeline automatically: parse, compress, encrypt, hash-chain, archive. Leave it running in the background.
Looks up archived logs by time range, service name, log level, or keyword, decrypts only the matching entries, and prints them back to you. It never has to download your entire archive to answer a question.
Walks the entire cryptographic hash chain, every chunk, in order, and confirms each one is exactly as it was when it was written. If even one byte anywhere was altered or deleted, this catches it and tells you exactly where.
The difference isn't features, it's who's actually in control of your data.
| AWS Glacier / Datadog / Elastic | ObsidianLog | |
|---|---|---|
| Who holds your data | The vendor's servers, under their access policies | Encrypted on your machine first; only you hold the key |
| Cost at scale | Billed per GB per month, teams often delete logs early to control cost | A fraction of the cost, plus ~90–97% compression on log data |
| Tamper-evidence | You take the vendor's word for it | Cryptographically provable, on demand, with obsidianlog verify |
| Lock-in | Proprietary storage and APIs, hard to migrate away from | Open source (MIT). Your data, your keys, exportable anytime |
ObsidianLog is planned as a three-phase project, exactly as proposed to the Sia Foundation.
The core pipeline, CLI, and Sia integration — shipped
Compression, encryption, hash-chaining
Vector-compatible HTTP endpoint
init, query, verify
Linux, macOS, Windows + Docker
Hosted by default, or bring your own
Full site with hands-on tutorials
Grafana and SIEM exports
Live on crates.io
Search, integrations, and a web UI — next
Full-text, fuzzy, anomaly detection
Plus Fluentd and Logstash
From S3 and CloudWatch
Plus a retention policy engine
Search, timeline, health dashboard
Sustaining the project, without gating the core — later
Pay in USD, no Siacoin needed
SLA, SSO, compliance support
Usage, cost, storage health
Splunk, IBM QRadar
MIT-licensed, always
If something's still unclear after this, that's worth telling us.
No, it sits alongside them. Datadog, Grafana, and ELK stay your "hot" tier for active, real-time monitoring. ObsidianLog is the "cold" tier: once logs aren't being actively watched anymore, they get archived somewhere cheap and permanent instead of deleted, and stay encrypted and verifiable so you can still trust and search them years later.
Sia is a decentralized cloud storage network: instead of one company like AWS owning the servers, your data is spread redundantly across many independent storage providers, encrypted so none of them can read it. ObsidianLog handles the Sia-specific parts for you. By default, obsidianlog init connects to a hosted service (sia.storage) that manages the underlying storage contracts, so you don't need to understand Sia's internals to get started.
Your logs are encrypted with AES-256-GCM on your own machine before anything is sent anywhere. The key is generated locally and stored in your OS keychain, it's never transmitted to Sia, to a storage host, or to us. Nobody but you can decrypt your archived logs, by design, not by policy.
Every log chunk is SHA-256 hash-chained to the one before it, a lightweight, private version of the same idea behind a blockchain. Run obsidianlog verify any time, and it walks the entire chain and tells you immediately if a single byte anywhere has been altered, deleted, or reordered.
ObsidianLog itself is free and open source (MIT license). You only pay for the underlying Sia storage you use, which is typically a fraction of the equivalent cost on AWS, Datadog, or Elastic, especially after compression, which usually shrinks log data by 90 to 97 percent.
No. You download a ready-to-run binary (or use the Docker image), run one setup command, and point your existing log shipper at it. There's no Rust toolchain to install, no smart contracts to write, and no crypto wallet to manage directly.
Yes. ObsidianLog can connect to any indexd deployment, including one you run yourself, if you want full control over every part of the stack instead of the hosted default.
Every decision, every bug, every fix is public. That's the point.
Read the code, fix something, or add a feature. Every PR is reviewed in the open.
Open-ended questions, ideas, or just want to see what others are building with it.
Tried it? Good or bad, real feedback from real use is what makes this better.
Set up takes under 15 minutes, and the local backend needs nothing but your machine to try.