Overview
AGLedger is change control for AI agents, bots and pipelines, delivered as a self-hosted signed ledger. Audit trail, approvals, agent memory and notifications: one API, one signed ledger. Agents run at full speed, and the trusted record builds itself as they work.
Audit trail. An agent records what it was asked to do, what it did and on whose authority, in one API call. Each record is signed and hash-chained the moment it is written. Work handed from agent to agent stays one traceable chain.
Approvals. Routine changes only need recording. The consequential few, such as a payment or a production change, wait for a verdict from the principal you designate: a person, your system of record, or a review agent you run. AGLedger holds the signed verdict.
Agent memory. Agents checkpoint their work as signed records. A fresh session picks up exactly where the last one stopped, in one query, with no prior conversation.
Notifications. Signed webhooks tell the systems and people around the work when an outcome lands. Settlement Signal webhooks carry verdicts to payment platforms.
Offline verification. Anyone holding the published public keys can verify the chain offline, with no access to your Server or to us. Built on open standards (COSE, in-toto, SCITT, Sigstore). The records map to the record-keeping provisions of the EU AI Act, ISO/IEC 42001 and NIST AI RMF.
Built for agents. Any agent, any LLM, any platform. The API tells an agent what to do next and how to fix a rejected call, so it gets it right on the first attempt. REST, TypeScript and Python SDKs, a CLI and an MCP server. Record types are your own JSON Schemas.
Federation. Servers in different business units, regions or companies share signed records peer to peer: the chain crosses the boundary, the data does not.
Deployment and licensing. A license is required for production use; this listing is the Enterprise license, required for an external database such as Aurora or RDS. Perpetual, per database instance. Helm on Amazon EKS or any Kubernetes, or Docker Compose on any Linux host. PostgreSQL is the only dependency. Runs air-gapped. No telemetry and no phone-home. Annual Support adds Updates and email support.
Highlights
- Audit trail: every change an agent records is signed and hash-chained in one API call. Anyone holding the public keys can verify the chain offline.
- Approvals: the consequential few wait for a verdict from a person or your own system, and AGLedger holds the signed answer. Signed webhooks tell everyone downstream.
- Yours to run: Amazon EKS with Aurora or RDS, any Kubernetes, or Docker Compose, even air-gapped. You hold the database, the keys and the records. No phone-home.
Details
Features and programs
Financing for AWS Marketplace purchases
Pricing
Dimension | Description | Cost/12 months |
|---|---|---|
AGLedger Enterprise - Single Database Instance | Perpetual Enterprise license for one AGLedger database instance, required for an external database such as Aurora or RDS. HA replicas included. Every feature: signed hash-chained records, approvals with verdicts and disputes, agent memory, signed webhooks, delegation, federation and custom record types. Security Fixes for supported versions included. Add Annual Support for Updates and support. | $3,000.00 |
AGLedger Annual Support | Annual support for one AGLedger database instance: all Updates, plus email support with initial response targets from 4 business hours for a production outage (P1) to 3 business days for a general question (P4). | $1,000.00 |
Dimensions summary
Top-of-mind questions for buyers
Vendor refund policy
Full refund available within 30 days of activation, no questions asked. Contact support@agledger.ai with your AWS account ID to request a refund.
How can we make this page better?
Legal
Vendor terms and conditions
Content disclaimer
Delivery details
AGLedger Container (Helm on EKS / Docker Compose)
- Amazon EKS
- Amazon ECS
Container image
Containers are lightweight, portable execution environments that wrap server application software in a filesystem that includes everything it needs to run. Container applications run on supported container runtimes and orchestration services, such as Amazon Elastic Container Service (Amazon ECS) or Amazon Elastic Kubernetes Service (Amazon EKS). Both eliminate the need for you to install and operate your own container orchestration software by managing and scheduling containers on a scalable cluster of virtual machines.
Version release notes
AGLedger 2.2.0 upgrades a 2.0.0 or 2.1.0 install in place and includes everything from 2.1.0. A record's contractVersion must name a registered version of its type, such as "1"; a value that names none is refused with 400 and the registered versions, where it used to bind the current one. A Server keeps a signed history of its own federation signing keys, served unauthenticated at /.well-known/agledger-federation-keys.json, so a receiver can verify messages sealed under an earlier key; change the key by booting the new one with the old one in AGLEDGER_FEDERATION_SIGNING_KEY_PREVIOUS. Co-sign counter-signatures are signed under their own domain, so co-sign between a 2.2.0 Server and an earlier peer fails until both run 2.2.0; other federation traffic is unaffected. A 401 from a peer is retried rather than failing the record. Records with no completion phase report gateMode, maxRevisions and maxDisputes as null. Dispute evidence is hash-bound on the signed chain, and audit exports can inline the filed evidence. Backups record the next transaction id instead of the WAL position, so restores from Aurora replay revocations. /openapi.json validates as OpenAPI 3.0.3. Each release adds one database migration (002 changes indexes only; 003 adds the federation key history), and a pod on the previous release keeps working against the new schema during a rollout or rollback.
Additional details
Usage instructions
Quick Start (Helm on EKS)
- Authenticate to the AWS Marketplace registry:
aws ecr get-login-password --region us-east-1 | docker login --username AWS --password-stdin 709825985650.dkr.ecr.us-east-1.amazonaws.com
- Put both secrets in files. Command-line arguments are readable by every local user through ps. The key must be base64 of PKCS#8 DER; base64 of the PEM text is a different thing and will not load.
umask 077 openssl genpkey -algorithm ed25519 | openssl pkey -outform DER | openssl base64 -A > vault-signing-key printf '%s' 'postgresql://USER:PASS@YOUR-DB-HOST:5432/agledger' > database-url
- Install with Helm (external PostgreSQL 17+ - Aurora, RDS, or self-managed):
helm install agledger oci://registry-1.docker.io/agledger/agledger-chart --version 2.2.0
--set image.repository=709825985650.dkr.ecr.us-east-1.amazonaws.com/ag-ledger/agledger
--set image.tag=2.2.0
--set-file database.externalUrl=./database-url
--set-file secrets.vaultSigningKey=./vault-signing-key
--set config.externalUrl=https://agledger.example.com
--set marketplace.productId=prod-gdyk7ehkopbnm
--set marketplace.serviceAccountAnnotations."eks.amazonaws.com/role-arn"=arn:aws:iam::YOUR_ACCOUNT_ID:role/agledger-marketplace
Notes:
- database.externalUrl: the migration role must be SUPERUSER (rds_superuser on RDS/Aurora); the schema installs an event trigger. For least privilege, point secrets.databaseUrlMigrate at a superuser URL used only for migrations and keep database.externalUrl DML-only.
- config.externalUrl: the public URL of this install - it is signed into every record as the issuer. Auto-derived from ingress.hosts[0] when ingress is enabled.
- marketplace.serviceAccountAnnotations: IRSA role with license-manager:CheckoutLicense and CheckInLicense. The Server checks your entitlement out and straight back in every 15 minutes while it runs, and once at boot unless a license key of your own is configured, which is read first. The call carries no record, agent or usage data and fails open: an unreachable License Manager never blocks anything.
- EKS worker nodes pull the image with their standard ECR read permissions once the account is subscribed.
- FIPS-mode clusters: FIPS providers carry no Ed25519, so write an ES256 key to the same file instead (openssl genpkey -algorithm EC -pkeyopt ec_paramgen_curve:P-256 | openssl pkcs8 -topk8 -nocrypt -outform DER | base64 -w0 > vault-signing-key) and add --set config.allowNonDefaultSigningAlg=true. Chain consumers need @agledger/verify 2.0.0 or newer. Federation is unavailable on FIPS hosts.
- Verify:
kubectl port-forward svc/agledger-agledger-chart 3001:80 curl http://localhost:3001/health
- Initialize the first organization and API key:
kubectl exec deploy/agledger-agledger-chart-api -- env NODE_OPTIONS= /nodejs/bin/node dist/scripts/init.js --non-interactive
Docker Compose (any Linux host)
After step 1 (docker login):
git clone --branch v2.2.0 https://github.com/agledger-ai/install.git cd install ./scripts/install.sh --image 709825985650.dkr.ecr.us-east-1.amazonaws.com/ag-ledger/agledger --version 2.2.0
Verify provenance (optional)
This image is a byte-for-byte copy of the linux/amd64 image of the signed public release agledger/agledger on Docker Hub, tagged there 2.2.0-amd64. Verify the public release with cosign + slsa-verifier (SLSA Build L3), then confirm this copy matches:
crane digest agledger/agledger:2.2.0-amd64 aws ecr describe-images --registry-id 709825985650 --repository-name ag-ledger/agledger --image-ids imageTag=2.2.0 --region us-east-1 --query imageDetails[0].imageDigest --output text
The two digests must be identical. Full steps: https://github.com/agledger-ai/install/blob/main/SECURITY.md
Documentation
More documentation:
- Quick start: https://agledger.ai/docs/quick-start
- Day-2 operations (backup, recovery): https://agledger.ai/docs/operations/day-2
- Full documentation: https://agledger.ai/docs
Resources
Vendor resources
Support
Vendor support
Annual Support covers troubleshooting, configuration and upgrades of the AGLedger Server and its Helm, Docker Compose and federation packaging, plus all Updates. Email support@agledger.ai . Initial response targets run from 4 business hours for a production outage (P1) to 3 business days for a general question (P4). A built-in script produces a sanitized diagnostic bundle you review and send to support. Docs and API reference: agledger.ai/docs.
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.