Open
respawn opens a job when work arrives in a channel it handles, such as a shared inbox or team chat, or when a person asks for it.
Home / AI software and development
AI that works the way your business already works — on a server you control.
respawn is a multi-user, multi-surface agentic harness for automating tasks. It primarily runs on frontier model APIs. Your data, workflows, tools, and secrets stay on hardware you own.
Local inference can be set up as well: certain tasks can use local models, or the entire harness can run on local models.
- Agentic: opens jobs, spawns sub-agents
- Primarily frontier model APIs
- Local inference for some tasks, or all
- Data, tools, and secrets on your hardware
Job 4127Vendor invoiceOpened by respawn · from the shared inbox
Open
- Check the senderrespawn · Your serverDone
- Check the invoice against the purchase orderSub-agent · Another frontier modelDone
- Enter the invoice in the books for reviewBookkeeper’s assistant · Your serverDone
- Confirm the deliveryVendor’s agent · Another companyDone
- Approve the paymentOffice manager · A personWaiting
The harness assigns each part and carries the work between them.
The difference
Someone has to open the chat.
A hosted chat can be pointed at your systems. ChatGPT, Claude, Grok, and the rest still wait for a person to open the window and ask. Until someone does, nothing moves.
respawn is the place in the middle. It takes work from one system, handles it by your rules, and puts the result in the next system. Your staff do not have to sit in a chat for that run to happen.
A hosted chatIdle
Waiting for someone to ask. The window waits. Until a person opens it and asks, nothing moves.
respawnRunning
From one systemBy your rulesInto the next system
- Shared inboxAn invoice is read and checked against the procedure the office already follows.The books, for review
- Team chatA question is answered from the business context, not from a blank chat.The same thread
- The work itselfA follow-up is drafted, and sent when your rules say to.The customer’s inbox
- Anything elseWork that needs a decision is handed to the person who makes it.The right person
Examples of the kind of work an office hands to respawn.
In the channels you already use
What it does for a business
Most offices already run on email, chat, and shared inboxes. Work piles up there — invoices, customer threads, vendor noise, follow-ups that someone meant to handle. respawn sits in those channels and carries the work through.
- 01
Reading what came in
Mail, chat messages, and requests that land in a shared inbox.
- 02
Following your procedures
The way your office already handles that kind of work.
- 03
Drafting or sending what should go out
A reply or an update, sent when your rules say to.
- 04
Handing the rest to the right person
Anything that needs a decision goes to the person who makes it.
Agentic
Jobs, other agents, and sub-agents
The harness is agentic. It can open jobs and coordinate with other agents, including agents from different companies. The harness is what assigns and carries that work. A person does not have to sit in a chat window for the run to happen.
An agent can spawn sub-agents that use other frontier models, and sub-agents that use local models. One job can draw on more than one model, and the business context stays on the server the business owns.
Assign
Each part goes to an agent: respawn itself, a sub-agent it spawns, another person’s assistant on the same server, or an agent from another company.
Carry
The harness carries the work between them. Nobody has to copy a message from one chat window into another.
Hand off
Anything that needs a decision goes to the right person.
An agent from another company stays that company’s agent. respawn coordinates with it; it does not become part of respawn.
Surfaces
Multi-surface, one harness
respawn is the layer underneath the places your staff already open every day. Same assistant, same memory of how you work — across the surfaces that matter.
Shared inboxes and day-to-day mail handled with your rules, then routed or answered instead of left unread.
Chat
Team chat where people already ask questions — answers that know the business, not a throwaway bot on a website.
The work itself
Tasks that need follow-through get done or handed off cleanly, so the queue moves instead of stalling overnight.
spawnpoint
The desktop and mobile client on top of that same harness. It is another surface, not a second product with a second memory. About spawnpoint
Teams
Built for teams, not a single operator.
Each person gets their own assistant, designed to work together on the same server — with shared business context, and without living inside one chat window.
An office is not one person in one chat. The office manager, the bookkeeper, the front desk, and the owner each have their own work, and each gets their own assistant. Those assistants work from the same business context on the same server, instead of from separate chat histories in separate accounts.
One serverOne office
OMOffice manager’s assistant
BKBookkeeper’s assistant
Shared business context
FDFront desk’s assistant
OWOwner’s assistant
Local models
Three ways the same harness can run
respawn primarily runs on frontier model APIs. Local inference is a real second path: an office can set up local models for certain tasks, or run the entire harness on them. Either way it is the same harness.
Frontier model APIs
The usual setup. The harness runs on hardware the business owns and uses frontier model APIs for the model work. Your data, workflows, tools, and secrets stay on hardware you own.
Frontier models
Local models
Frontier models: the model work. Local models: not set up.
Local inference for certain tasks
Local inference is set up for certain tasks. Those tasks run on local models, and the rest of the work uses frontier model APIs. The business context stays on the server the business owns.
Frontier models
Local models
Frontier models: the rest of the work. Local models: certain tasks.
Local inference for the entire harness
The entire harness runs on local models. The harness, the models, and the business context are all on hardware the business owns.
Frontier models
Local models
Frontier models: not used. Local models: all of the model work.
In every setup, this stays on the server
- Data
- Workflows
- Tools
- Secrets
Ownership
Yours to keep
Subscription AI products hold the memory of how you work. When the plan changes or the vendor goes away, that history goes with them. respawn is different: it runs on a machine you own, so the business context stays with the business.
Local by design.
Conversations, procedures, tools, and secrets live on your server — not in a vendor cloud you rent month to month.
No vendor lock-in.
You are not dependent on someone else’s roadmap, pricing, or uptime. If a service disappears, your automation and its memory are still yours.
Swap the brain anytime.
Reasoning-engine agnostic — swap the model when a better one appears. The harness, the workflows, and everything it has learned can respawn again without restarting from zero.
The email trust layer
Email the model is allowed to read
A chat product connected to your invoices and QuickBooks, with an address anyone on the internet can write to, will treat whatever arrives as something to act on. That includes a stranger, a spoof, and a real mailbox that has been taken over. respawn does not.
Words in an email are data. They are never instructions.
From “A company you work with”Not from who it claims to be
Who it claims to be
respawn includes a hardened email layer that makes sure a message is from who it claims to be. Mail that only pretends to come from a person or a company is not treated as theirs.
From Someone newNot a trusted sender
Strangers and phishing
Mail from someone the business does not already trust, and mail built to deceive, cannot reach the books or become an order to the model.
From A real customer’s mailboxReal sender, still only data
A real mailbox, taken over
A real sender whose mailbox has been broken into can still send mail that looks legitimate. That message still cannot move what is in the books, and the words in it are still not orders.
Also from Tech Direct Services
Websites and dashboards
Websites
From a plain company website to a dynamic site built around your AI.
Dashboards
A dashboard for the metrics and the systems your business runs on. Sales, jobs, inventory, calls, and tickets on one screen, read from the systems you already use.
Who makes it
Designed from scratch by Tech Direct Services
We built respawn for the businesses we already support — the same customers who call when the network breaks or when mail needs to keep moving after hours. It is not a bolt-on product from somewhere else. The harness was written from the ground up in Go so it can be hosted on your own hardware, operated by people you trust, and shaped around how each office actually runs.
If you already work with Tech Direct Services, this is the automation layer behind the work we do for you.
AI that works the way your business already works — on a server you control.respawn is a multi-user, multi-surface agentic harness by Tech Direct Services. Your data, workflows, tools, and secrets stay on hardware you own.
respawn.bot
Ask us about respawn for your office.
Call (888) 261-2258, or use Chat With Us in the corner.