Why not just use print() or logging?

Logging shows fragments. Tether connects the prompt, action, file diff, command output, failed test, and recovery point in one local execution graph.

Can I use this with production code?

Yes. It's a local proxy, so your real code doesn't change. Use it locally for debugging, or keep it running. Tether only stores traces locally and never sends anything anywhere.

How much money does caching actually save?

It depends on your agent. If you're iterating on prompt logic and re-running the same retrieval steps, caching saves you 50-90% of API spend while you debug. Each cached hit costs $0.0000.

Does Tether add latency to my agent?

Negligible. Tether runs locally on your machine. The only overhead is the proxy hop, which is less than 1ms. Real LLM calls are the bottleneck, not Tether.

Can I share traces with my team?

Not yet. Each developer runs their own Tether instance locally. Export as JSON is coming in a later release.

How does Tether intercept LLM calls without SDK changes?

Tether runs a local HTTP proxy on your machine. You point your AI client's base_url at http://localhost:8080/v1 — that's the only change required. Tether forwards requests to the real provider and records everything locally.

Does Tether send my prompts or API keys to the cloud?

No. Tether is fully air-gapped. Your prompts and responses stay on your machine. macOS secrets use Keychain-aware storage, Linux uses the same local proxy model, and traces stay local.

Which LLM providers does Tether support?

Tether supports OpenAI, Anthropic, Ollama, LM Studio, and any provider that accepts an OpenAI-compatible base_url. It also works with LangChain, LangGraph, LlamaIndex, and similar frameworks.

Yes. Tether is free during the alpha period. The core proxy is open source.

Trust starts at the machine

Security

Tether reduces debugging risk by keeping the recovery loop local, explicit, and inspectable.

Contact security Back to product

01Local evidence, not hosted dashboards

02Threat model for the alpha

Local evidence, not hosted dashboards

Tether is built for debugging sensitive agent runs without turning every prompt into another SaaS data copy. The core security claim is narrow: local traces, explicit provider calls, platform-local secrets, and inspectable replay state.

Trace database stays on the developer machine
Provider traffic follows your configured endpoints
No hosted Tether workspace is required for local debugging

Threat model for the alpha

The alpha should be evaluated as a local debugging tool, not as a compliance platform. The important questions are whether prompts are copied elsewhere, whether secrets sit in project files, and whether replay changes can be traced back to a concrete node.

No claim of SOC 2 or enterprise audit coverage yet
Local machine compromise remains out of scope
Provider-side data handling is governed by the provider you choose

Local trace artifacts100%

Prompt, response, cache, replay, and metadata records are designed for local SQLite storage.

Required hosted Tether services0

Local debugging does not require a Tether account, cloud dashboard, or remote project sync.

External provider boundaryExplicit

Only model-provider requests leave the device, and they follow the provider endpoints you configure.

Compliance postureAlpha

Security posture is architecture-led today; formal certifications are not claimed.

Data flowWhat moves, where it lands, and where Tether stops

01Your app

Sends OpenAI-compatible requests to localhost.

02Local proxy

Captures request metadata and forwards the call.

03Provider

Receives only the configured model request.

04Local SQLite

Stores trace, cache, replay, and failure evidence.

05Inspector

Reads local records for debugging and replay.

Data boundary matrixConcrete storage and egress rules

AssetStored atLeaves device?Control

PromptsLocal trace databaseOnly inside configured provider requestsRequest-level inspection

Model responsesLocal trace databaseNo Tether-hosted uploadReplay diff and node history

API keysPlatform-local secret storageUsed only to call chosen providersOS and local proxy secret handling

Cache entriesLocal cache storeNo remote cache syncCache hit metadata

Replay editsLocal replay stateNo background telemetry pathInvalidation and downstream scope

Risk registerWhat is covered, what is not, and who owns it

Owned by user machine

Local data at rest

Use OS disk encryption and keep workspace access tight.

Developer

Provider-dependent

Provider retention

Choose provider settings and contracts that match the data sensitivity.

Configured provider

Not claimed

Formal compliance

Add published audit artifacts only when the product surface is stable.

Tether roadmap

Requires deliberate action

Trace export

Review exported files before sharing them outside the machine.

Developer

QuestionAnswerProof

Prompt and response storageTrace data is written to local SQLite on the developer machine.~/.Tether/traces.sqlite

API key handlingProvider secrets use platform-local storage paths instead of plaintext project files.macOS Keychain and local proxy settings

Telemetry postureThe app does not require a hosted Tether account for local debugging.No cloud workspace

Reproducible failuresFailed runs can be replayed from a selected node with mocked outputs.Replay chain