Guide
Source Code Escrow: When Founders Need It and How to Set It Up
What source code escrow is, when it protects a founder working with an outside dev team, what to deposit, which release conditions to negotiate, and how to check costs.
By Vendor Stack Review · Reviewed November 17, 2026
The short answer
- Source code escrow is a three-party arrangement: the software supplier deposits its source code and build materials with an independent agent, who releases them to the customer if agreed events happen, such as the supplier going out of business or stopping support.
- If your contract assigns the IP to you and the code already lives in a repository your company owns, escrow adds little. Fix ownership and access first.
- Escrow earns its cost when a vendor keeps something you depend on: a proprietary framework it licenses to you, a platform it hosts for you, or a product your operations can't run without.
- A deposit nobody has verified is only the supplier's word that it is complete. Pay for at least a build check on anything you would actually need to recover.
Most founders meet source code escrow in one of two ways: a lawyer suggests it while reviewing an agency contract, or an enterprise customer asks for it during procurement. In both cases it sounds like the safe choice. Sometimes it is. Often it is a paid answer to a problem the contract should already have solved.
This guide is for founders and operators who rely on an outside engineering team, studio or software vendor. It covers how escrow works, the situations where it protects you and the ones where it doesn't, what a useful deposit contains, the release conditions worth negotiating, and how to check what it will cost.
How source code escrow works
An escrow agreement has three parties. The depositor is the supplier who wrote or owns the software. The beneficiary is the customer who depends on it. The escrow agent is an independent company that stores the deposit and follows the agreement's rules for releasing it.
- The parties sign an agreement, usually the agent's standard form with negotiated changes, that names the software, what the deposit must contain, how often it is updated and which events trigger a release.
- The supplier deposits the source code and supporting materials, then updates the deposit on the agreed schedule, typically with each release or at a fixed interval.
- If a release event happens, the beneficiary applies to the agent. The agent notifies the supplier, who has a set period to dispute the claim under the agreement's procedure.
- If the claim stands, the agent releases the deposit, and the beneficiary uses it under the rights the agreement or the main licence grants.
Two details decide whether this is worth anything. First, the deposit has to be complete enough for a different team to build and run the software. Second, the beneficiary needs the legal right to use, modify and maintain the released code; a release without those rights hands over files you may not be allowed to touch. Check both against the agreement and your main contract, not the agent's marketing.
Do you need it? It depends who holds the code
Escrow protects a customer who does not hold the source code. So the first question is not "should we use escrow" but "what would we be missing if this vendor disappeared tomorrow".
| Your situation | Is escrow useful? | Do this first |
|---|---|---|
| An agency or studio builds a custom product for you, the contract says it "hereby assigns" the IP, and the repository sits in your company's account | Rarely. You already hold the code and own it. | Confirm the assignment wording and that every repo, cloud account and domain is in your company's name. |
| The vendor builds on its own framework, platform or libraries and licenses those parts to you | Yes, for the licensed components you can't replace quickly. | Negotiate the licence scope, then escrow what stays with the vendor. |
| The vendor hosts and runs the product, and you don't control the cloud accounts | Yes, as SaaS escrow covering deployment materials and environment access, not just code. | Move hosting accounts into your name where you can; escrow what you can't. |
| You license an off-the-shelf product your operations depend on from a small supplier | This is the classic case for escrow. | Ask whether the supplier already has a multi-beneficiary agreement you can join. |
| A freelancer is building your MVP | Usually overkill. | Put the code in your own repository from the first commit and get a written IP assignment. |
| An enterprise customer asks your startup for escrow | You are the depositor here, not the beneficiary. | See "When a customer asks you for escrow" below. |
The first row is where founders most often overspend. Escrow cannot fix a weak ownership clause: if the contract only licenses the code to you, a released deposit comes with the same limits. Read IP assignment vs license in custom software agreements before paying for an escrow account, and exit planning and codebase handover for the account controls that make a handover possible without one.
What a useful deposit contains
"The source code" is not enough to rebuild a modern product. Write the deposit contents into the agreement as a list, so the agent can check against it:
- Source code for every component you depend on, including mobile apps, background workers and internal tools, not just the main application.
- Build scripts, dependency manifests and lockfiles, and the exact versions of languages and runtimes used.
- Infrastructure-as-code and deployment configuration: whatever the supplier uses to create servers, queues, storage and networking.
- Database schemas and migrations, and seed or reference data the product needs to start.
- A list of third-party services, APIs and licensed libraries, with the account or licence each one runs under.
- Documentation a new team needs: architecture overview, build and deploy instructions, and the names of the environment variables and secrets the system expects.
- For hosted products: how credentials and environment access are handed over at release, and where backups of your data are kept.
Agree how often the deposit is refreshed. A deposit made at launch and never updated describes software you no longer run. Tying updates to each production release, or to a fixed monthly or quarterly schedule, keeps it current; ask whether the agent can take deposits automatically from the supplier's repository host.
Release conditions worth negotiating
Release conditions are where escrow agreements succeed or fail. Vague triggers invite disputes at exactly the moment you need the code. Common conditions include:
- Insolvency, bankruptcy or the supplier ceasing to trade.
- Failure to support or maintain the software as the support agreement requires, after written notice and a defined cure period.
- The supplier discontinuing the product or the version you use.
- A sale or change of control after which the new owner does not take on the support obligations.
- Material breach of the support or service terms that is not cured in time.
For each trigger, the agreement should say what evidence the beneficiary provides, how long the supplier has to dispute it, who decides a dispute, and what the beneficiary may do with the code once it is released. Insolvency triggers also interact with insolvency law, which differs by country; in the US, lawyers usually look at section 365(n) of the Bankruptcy Code, which concerns licensees' rights when a licensor goes bankrupt. Have counsel who handles technology contracts review these clauses alongside your termination and transition terms.
Verification: the part people skip
Most escrow agents offer levels of verification, priced separately from the basic agreement. Names differ between agents, but they generally run from a light check to a full rebuild:
| Level | What the agent checks | What it tells you |
|---|---|---|
| Deposit only | That a deposit arrived and can be stored | Almost nothing about whether it is usable |
| Inventory or file check | That the files are readable, not corrupted, and match the listed contents | The supplier deposited roughly what it said it would |
| Build verification | That the code compiles or builds into working artefacts using the deposited instructions | Another team could produce the software from the deposit |
| Full functional test | That the built software deploys and runs as expected in a test environment | The closest you get to proof of recovery before you need it |
Match the level to what is at stake. For a component whose loss would stop your business, a build check is the minimum worth paying for, and a repeat check after major changes keeps it honest. The same habits that make a vendor's repository easy to audit make its deposit easy to verify; see GitHub and repository signals that reveal vendor quality.
Single-beneficiary, multi-beneficiary and SaaS escrow
- Single-beneficiary (often called three-party) agreements are negotiated between one supplier, one customer and the agent. They suit custom work and give you the most control over triggers and deposit contents.
- Multi-beneficiary agreements are set up once by the supplier on the agent's standard terms, and each customer joins as a beneficiary. They are quicker and cheaper to join, but you take the terms as written.
- SaaS escrow covers hosted software, where the code alone is not enough. Depending on the agent, it can include deployment materials, cloud account access, data backups and arrangements to keep the live environment running for a period after a release.
What it costs, and who pays
Escrow agents usually charge per agreement, often as a yearly fee, with verification and extra beneficiaries priced on top. Fees change and depend on the agreement type and verification level, so check the agent's current price list or ask for a quote rather than relying on a figure from an article.
Who pays is negotiable. Because the arrangement protects the customer, the customer often pays, but a supplier that wants enterprise deals may carry the cost of a multi-beneficiary agreement itself. Settle it in the contract, along with who pays for verification and for a release.
When comparing agents, look at their standard agreement, the verification levels on offer, whether they handle SaaS deposits, where deposits are stored, and how their release and dispute process works. Escode, EscrowTech and Codekeeper are among the agents that publish this information; we have not tested or ranked them, and naming them is not an endorsement.
When a customer asks you for escrow
Founders selling software to larger companies are often asked to put their own code in escrow during procurement. Here you are the depositor, and the risks run the other way:
- Offer a multi-beneficiary agreement on an agent's standard terms, so each new customer joins it instead of negotiating its own.
- Push back on loose triggers such as "failure to support" without a notice and cure period; a disputed release can put your code in a customer's hands over a missed ticket.
- Limit what the customer may do with released code to maintaining its own use of the product.
- Automate deposits from your repository so keeping them current costs no engineering time.
- Keep documentation in a state you would be comfortable having verified. Acquirers ask similar questions; see technical ownership models for founders ready to sell.
A checklist before you sign
- Write down what you would be missing if the vendor stopped answering tomorrow: code, accounts, data, licences, knowledge.
- Fix what the contract can fix: IP assignment, repository ownership, cloud and domain accounts in your company's name.
- Escrow only what remains with the vendor, and name it in the agreement.
- List the deposit contents, including build, infrastructure and documentation materials.
- Set the update schedule and how deposits are made.
- Define each release trigger with evidence, notice, cure and dispute steps.
- Confirm your rights to use, modify and maintain released code.
- Choose a verification level, and when it is repeated.
- Agree who pays for setup, renewals, verification and release.
- Put a reminder in the calendar to check the deposit is still being updated.
Common questions
Is source code escrow worth it for an MVP built by an agency?
Usually not, if the contract assigns the IP to you and the code is in your own repository from the start. That costs nothing and gives you more than escrow would. Escrow becomes worth considering when the agency keeps parts of the stack it licenses to you, or hosts the product for you.
Is read access to the vendor's GitHub repository the same as escrow?
No. Access can be revoked, and a repository alone may not include build, infrastructure and deployment materials. A repository owned by your company, with the vendor as a collaborator, is stronger than escrow for custom work; read access to the vendor's repository is weaker than both.
Can escrow replace an IP assignment clause?
No. Escrow controls when you get the code; the contract controls what you may do with it. A release under a narrow licence leaves you with the licence's limits.
How much does source code escrow cost?
It depends on the agent, the agreement type, the number of beneficiaries and the verification level. Agents publish price lists or quote on request; check current figures with at least two before you budget.
This guide explains how these arrangements usually work; it is not legal advice. Have a lawyer who handles technology contracts review any agreement before you sign it.


