SOFTWARE RECOVERABILITY
Castler independently rebuilds, deploys, and runs your vendor’s software in a clean environment—then a named engineer signs the evidence. Every verified release becomes a current, examinable Proof of Recovery for your board, regulator, insurer, and recovery team.
Coverage can include source code, SaaS, AI models, data, and IP.
LIVE RECOVERY RUN
STREAMINGcore-banking-platform · v9.4
java/spring · postgres · kubernetes
WHAT IS SOFTWARE RECOVERABILITY
Software recoverability is the independently verified ability to rebuild, deploy and operate a vendor's software application without the vendor's involvement. It is distinct from software escrow, which stores code, and from backup, which copies data. Recoverability requires an independent rebuild in a clean environment, a successful deployment, and a signed engineer's confirmation that the system runs. Castler SRP is the only platform that produces this evidence as a standardised, signed artefact.
THE PROBLEM THIS SOLVES
An escrow agreement says that if a trigger event occurs — vendor insolvency, acquisition, or discontinuation — the source code will be released and you will be able to get the software running again.
It does not say the code is complete. It does not say the build environment is documented. It does not say a single person alive knows how to deploy it. It does not say the deposit has been updated since the original filing. It does not say the code has ever compiled in any environment other than the vendor’s own.
The gap between an agreement and a working recovery has historically been measured in months. For complex enterprise applications, that gap is often permanent — not because the will to recover is absent, but because the knowledge required to do it has never been captured.
Castler closes that gap. Not with better contract language. With a repeatable, tested, signed recovery procedure that runs on every release and produces examinable evidence every time. It begins at vendor onboarding, when the first deposit, custody workflow, and verification schedule are established — not as a regulatory retrofit years later.
FAILURE-MODE COVERAGE
Distinct glyphs preserve meaning without relying on colour alone.
Deposit is corrupt, encrypted without keys, or unreadable
Cloud Storage · Partially detectedStandard Verification · Detected before you need itPremium Recoverability · Detected before you need itRepository incomplete, source files or modules missing
Cloud Storage · Partially detectedStandard Verification · Detected before you need itPremium Recoverability · Detected before you need itDependencies undeclared or held in private vendor repositories
Cloud Storage · Discovered during outageStandard Verification · Detected before you need itPremium Recoverability · Detected before you need itSource code does not compile, no usable build can be produced
Cloud Storage · Discovered during outageStandard Verification · Detected before you need itPremium Recoverability · Detected before you need itBuild instructions absent, incomplete or incorrect
Cloud Storage · Discovered during outageStandard Verification · Detected before you need itPremium Recoverability · Detected before you need itUndisclosed third-party or open-source licence exposure
Cloud Storage · Discovered during outageStandard Verification · Detected before you need itPremium Recoverability · Detected before you need itBEFORE AND AFTER
WHAT CASTLER PRODUCES
For every critical vendor deposit, Castler’s agentic verification engine rebuilds the software in a clean environment, deploys it, replicates the declared production architecture, and generates a structured set of audit-grade artefacts. A named Castler verification engineer reviews and signs the result.
The full pack is delivered as a sealed, numbered Proof of Recovery certificate. It is examinable, linked to the exact release tested, and refreshed when the vendor ships a new version.
Castler · Certificate of Recoverability
Aurora Core · Release 24.11.3
A. Mehta, Verification Engineer
castler.io/verify/POR-2026-04821
A complete log of the build process: compiler, dependencies, build tools, errors encountered and resolved, and final binary outputs. It is evidence that the code compiles from the deposit alone rather than from undocumented knowledge held only by the vendor.
A step-by-step operational guide for deploying the application in your environment if a trigger event occurs. It is written so your team can execute it without the vendor present, including environment setup, configuration, health checks, and rollback instructions.
Evidence that the declared production architecture — database, network topology, service configuration, and environment variables — has been replicated and validated. This proves recovery extends beyond compilation to an operable system.
A complete inventory of every dependency, library, runtime, and component included in the verified release, formatted to SPDX and CycloneDX standards. The SBOM creates an auditable software-supply-chain record for the exact release tested.
A numeric score from 0–100 representing verified recoverability confidence for the release. It reflects build completeness, dependency resolution, deployment success, replication fidelity, and the remaining exceptions that require action.
A numbered, dated signature from a named Castler verification engineer. It is not a probabilistic output: it is an auditable artefact with an accountable professional behind the claim and a public verification reference.
HOW IT WORKS
The first verification establishes the recoverability baseline. Subsequent release cycles reuse the mapped architecture and run automatically, so evidence remains current without manual scheduling.
Your team deposits the source code, build inputs, runtime configuration, and environment documentation into Castler’s secure cloud custody. Every subsequent release is captured automatically through repository integration, so the materials used for verification remain aligned with the production version.
Castler’s agentic verification engine attempts to rebuild the software from the deposit alone, without the vendor present. AI agents map dependencies, resolve toolchain requirements, compile the application, and generate the full SBOM. Missing information is flagged and routed into a structured reconciliation process.
Castler deploys the rebuilt application in a clean cloud environment and replicates the vendor’s declared production architecture: database schema, service mesh, network configuration, environment variables, and runtime services. The deployment is validated against the declared specification.
A named Castler verification engineer reviews the build report, deployment runbook, replication report, SBOM, confidence score, and exceptions. The resulting Proof of Recovery is numbered, dated, signed, and delivered to your account with a verification reference.
Initiated on demand, delivered in hours, re-run on every release
THE ENGINE
Each stage converts vendor-held knowledge into a documented, repeatable recovery capability that your team can execute independently.
Castler ingests the deposit and inventories every dependency, runtime, build toolchain, repository, and environment variable declared by the vendor. Missing dependencies are flagged immediately. The map becomes the control record for every subsequent stage.
Where the deposit is insufficient to build, the gap is recorded as an exception rather than a failure. Most gaps are closed from the deposit and repository history alone. Where a gap cannot be closed independently, Castler raises it with the depositing vendor once, and the recovered detail becomes permanent documentation rather than vendor-held memory. Every subsequent verification runs without vendor involvement.
Castler authors the Emergency Deployment Runbook for your team, not the vendor’s. It covers environment setup, database initialisation, configuration management, health checks, operating dependencies, rollback procedures, and the order in which the recovered system must be brought online.
AI agents compile the application from the deposit in a clean environment. The complete build log and SBOM are captured, every resolved dependency is recorded, and the confidence score is recalculated after each successful or failed build run.
The rebuilt application is deployed on Castler’s cloud infrastructure. Automated health checks validate service availability, API response, database connectivity, and declared operating behaviour. The runbook is executed as written and any gaps are recorded for correction.
The verified application is re-deployed on your declared production environment or a representative replica. Parity between the Castler deployment and the target architecture is confirmed. This is the final validation that recovery — not merely build — is achievable.
TECHNICAL COVERAGE
Castler verifies the software your organisation depends on across languages, repository structures, cloud targets, and orchestration systems. Complex or proprietary toolchains are assessed during deposit mapping so scope, dependencies, and exceptions are known before recovery is ever required.
Java, Kotlin, Python, Go, Node.js, TypeScript, .NET, C#, Ruby, PHP, Rust, and Scala. Framework support includes Spring, Django, FastAPI, Express, Next.js, .NET Core, and Rails. Custom build toolchains are assessed during deposit mapping.
Mono-repo, multi-repo, submodule trees, private registries, and Git. Binary and compiled-only deposits are assessed separately, with exceptions and evidence requirements made explicit before verification begins.
AWS, Google Cloud Platform, and Microsoft Azure. Other cloud targets are assessed case by case against the architecture declared by the vendor and the target environment selected by the beneficiary.
Docker and OCI-compliant images, Kubernetes across EKS, GKE, AKS and on-premises clusters, Helm charts, and Terraform.
Hours
Per verification
~90%
Lower verification cost
100%
Critical-Software Estate coverage
Per release
Evidence stays current
PRICING TIERS
All tiers use the same secure cloud-custody foundation. They differ in verification depth and how far the Proof of Recovery cycle goes from build evidence to architecture mirror testing.
Tier 1
$1,000 / application / year
Secure, versioned, access-controlled custody of source code, build inputs, and runtime configuration. Every release is captured automatically with a full audit trail and encryption in transit and at rest.
WHAT YOU PRODUCE
A current, versioned deposit — the foundation for verification.
BEST FOR
Organisations beginning a recoverability programme, or vendors meeting customer escrow requirements.
Tier 2
$2,500 / application / year
Everything in Cloud Storage, plus agentic build verification, SBOM generation, dependency-resolution reporting, a build confidence score, and an Engineer’s Seal on every verified release.
WHAT YOU PRODUCE
A signed Proof of Recovery covering build completeness and SBOM accuracy.
BEST FOR
Institutions meeting baseline RBI, SEBI, IRDAI, DORA, or customer assurance requirements.
Tier 3
$5,000 / application / year
Everything in Standard Verification, plus deployment-runbook authoring, replication on Castler cloud, mirror testing against the declared production architecture, and the highest-confidence engineer seal.
WHAT YOU PRODUCE
A complete Proof of Recovery: build, deploy, replicate, and mirror — signed and numbered.
BEST FOR
Board-level mandates, insurer underwriting evidence, and deep-verification requirements.
Volume ladder: Essential, Silver, Gold, and Platinum coverage levels are available within each product tier for multi-vendor estates.
See full pricing →Book a briefing to discuss which tier fits your Software Estate →WHO THIS IS FOR
CIO or CISO accountable for vendor risk and operational resilience
Head of IT Risk or Third-Party Risk Management
Compliance teams responding to RBI, SEBI, IRDAI, DORA, or FCA requirements
Boards and risk committees requiring current recovery evidence
ISVs and SaaS companies contractually required to establish escrow
Technology companies using recoverability as an enterprise trust signal
Vendors managing multiple customer escrow obligations through one platform
REGULATION MAPPING
Every artefact in the Proof of Recovery pack is designed around a named operational obligation. The mapping below shows which evidence supports each requirement; individual regulation pages explain the official text, scope, status, and compliance timeline.
| Regulation | Article / section | Castler artefact |
|---|---|---|
| RBI IT Directions | §12(f) | Deposit + Build Report + Engineer’s Seal |
| RBI PSO Directions | §17(c) | Deposit + Build Report + Engineer’s Seal |
| SEBI CSCRF | Vendor continuity clause | Full Proof of Recovery pack |
| IRDAI CS Guidelines 2023 | Software escrow clause | Deposit + Build Report |
| EU DORA | Articles 9(2) and 28 | Full Proof of Recovery + Replication Report |
| UK PRA SS2/21 | Operational resilience evidence | Proof of Recovery + Deployment Runbook |
| MAS TRM | Technology risk management | Full Proof of Recovery pack |
| APRA CPS 230 | Operational risk and service providers | Full Proof of Recovery pack |
SOFTWARE RECOVERABILITY
The Proof of Recovery exists before the incident, not after it. Book a 15-minute briefing to see how Castler maps to your Software Estate and regulatory obligations.
Book a 15-min briefing