VS Code for the editor. Desktop when you want full control.
Same local pipeline and license. Two ways to use it.
Direct VS Code experience
VS Code extension
The direct VS Code experience. Agent and Ask chats live in an editor tab. Pick the provider you already use, watch Live Traffic and cost beside your files, and approve tools before they write. One .vsix — no admin rights.
Full control
Windows desktop
Full control on the same pipeline. Session timeline, Sandbox to try a change before it hits real chats, Privacy Audit to prove prompts are not stored, every dashboard and setting, plus the wire CLI. Optional — nothing in the extension requires it.
See what Wire does
Every chat from the extension — and every request from the wire CLI or an
SDK on the local /v1 door — runs through the same local pipeline. Pick a goal, then flip through Live Traffic and the stats
for Guard, Terran, Router, Compress, and Advise. Mutation stages stay off until you choose.
Observe visibility is included in the trial, then $39 to keep — Observe is not
free.
Pick what you want Wire to do
Start with visibility, protect secrets, spend less, prefer local models, or coach your prompts. Traffic from your editor can pass through Guard, Terran, Router, Compress, and Advise — those stages stay off until you choose. Observe visibility is included in the trial, then $39 to keep.
Install the VS Code extension
Download the .vsix and install it with Extensions: Install from VSIX. The extension finds or starts the local pipeline — no admin rights, certificate, or driver. Later builds arrive through the chat Update button. Everything runs on loopback, not a cloud analytics service.
Chat and watch live traffic
Start an Agent or Ask chat in an editor tab and pick the provider you already use — Cursor, Copilot, Claude, OpenAI, Anthropic, Ollama, or Gemini. No lock-in. Every turn runs through the local pipeline; Live Traffic and spend estimates update beside your code — prompts are never stored. Add Guard, Router, Compress, Terran, or Advise when you want them.
Add the Windows desktop for full control
Optional. The installer is the desktop command center: session timeline (including compare), Sandbox, Privacy Audit, complete Settings, and the wire CLI. Same license as the extension. Per-user install; no certificate, driver, or service.
Then explore traffic, each stage, and cost
These are real product screens. In VS Code you chat and open traffic, cost, and pipeline boards beside your code. On Windows, the desktop app is the same pipeline with full control — Sandbox, Privacy Audit, and the rest of the settings. Sanitized telemetry only, prompts never stored.
VS Code — Agent chat
Agent and Ask chats live in an editor tab. Approve tools, pick a provider, and every turn runs through the local Host — no admin rights, certificate, or driver.
Why it matters: Chat where you already work — you do not have to open a separate app first.
VS Code — Provider picker
Switch Cursor, Copilot, Claude, OpenAI, Anthropic, Ollama, or Gemini per chat — or an OpenAI-compatible endpoint. No vendor lock-in.
Why it matters: Use the models you already pay for, or local Ollama, without switching editors.
VS Code — Live Traffic
Live Traffic overview docked in the editor — traffic mix, tokens, latency, and spend estimates. No prompt text displayed or stored.
Why it matters: See the pipeline next to your code without opening a separate window.
VS Code — Cost Estimation
Projected cycle spend and cost-by-model from observed tokens, in the same Light VS Code window.
Why it matters: Spend estimates stay beside the chats that created them.
VS Code — Traffic & routing
Token flow, requests by project, estimated Router savings, and provider mix — in the editor, beside the chats that created them.
Why it matters: See what your agents are doing and what it costs without leaving VS Code.
VS Code — Compression savings
Tokens saved per chat, requests compressed, and token/dollar savings by provider — docked beside your code.
Why it matters: Quantify compression in dollars from the same window you chat in.
VS Code — Pipeline health
Guard blocks, Router matches, Compress savings, and Terran assess-to-enrich — one board in VS Code.
Why it matters: Monitor Guard, Router, Compress, and Terran beside your code.
VS Code — Waste & idle burn
Idle-burn minutes, tokens, and spend for the cycle, plus per-IDE idle detail — in the editor, not a separate app.
Why it matters: Catch the drip while you still have the chat that caused it open.
Aura Guard — Working samples
Canned prompts for inline secrets, config/env pastes, and protected concepts — run through the live pipeline and see block/redact outcomes.
Why it matters: Catch accidental credential paste before secrets leave your machine or hit a cloud API.
Aura Router — Stats
Route matches, tokens by route, before/after provider and model mix, and delegated volume — including Copilot CLI and cheaper-model downshifts.
Why it matters: Quantify routing wins in tokens and mix shift, not just a toggle.
Aura Advise — Advice
Prompt quality meter (clarity / structure / discipline) plus live advice cards from your own traffic — rule keys and counts only, never prompt text.
Why it matters: Catch vague prompts, retry storms, and costly habits as they form — without storing what you typed.
Privacy Audit
Built-in canary test injects unique markers through the live pipeline, then scans local storage — every location comes back clean.
Why it matters: Verify the no-content-storage model — do not just take our word for it.
Sandbox
Multi-chat composer to exercise Guard, Router, Compress, and Terran in dry-run (default) or optional live mode — without configuring an IDE.
Why it matters: Try before you proxy — validate pipeline behavior safely without touching editor settings.
Aura Compress — Multimodal
Right-size oversized inline images and optionally distill screenshot-like frames to text with local vision — remote image URLs are never fetched.
Why it matters: Vision turns stop burning tokens on full-resolution screenshots you never meant to re-send every turn.
Aura Terran — Connection
Terran injects retrieved chunks into chat requests. Point Wire at a local REST URL, pick namespace and hybrid search, and watch the IDE gRPC injection map (Cursor StreamChat and related backends) flip to Ready — fail-open when Terran is offline.
Why it matters: Index once in Terran; enrich every proxied chat in Wire without changing IDE base URLs.
Aura Terran — Working samples
Canned prompts for policy lookup, runbook search, and API reference — preview the corpus chunks Terran would inject before upstream forwarding.
Why it matters: Validate local RAG injection on realistic questions before routing real editors through Wire.
Aura Advise — History
Every advice card Advise has issued — active, snoozed, and dismissed — with sticky filters and CSV/XLSX export.
Why it matters: Review coaching over time without losing dismissed tips or replaying stale advice next to “Waiting for samples.”
Privacy and data
Aura Wire is local-first software. It does not upload your prompts to Halo Softworks or operate a cloud analytics service for your coding sessions.
- Runs locally on your machine — not a Halo Softworks cloud analytics service.
- Sanitized telemetry only: provider, model, tokens, latency, and cost estimates.
- Prompts, responses, and message text are never persisted.
- 7-day full-pipeline trial starts with your first chat — no card. After trial, keep Observe for $39 or buy the full pipeline.
Read our Privacy Policy, EULA, and Terms of Sale for website and licensing details. Questions? Contact Halo Softworks.
Stronger together
Index once in Terran (sold separately). Wire’s Terran stage is a bridge that injects that corpus into every Wire chat — or keep using Terran passthrough and MCP when that fits better.
- Aura Wire alone — Chat from VS Code, the wire CLI, or any /v1 client. Wire observes, protects, routes, compresses, and coaches — without Terran and without a corpus.
- With Aura Terran — Wire’s Terran stage is a bridge to a separately purchased local Terran install: ranked chunks enter the pipeline (fail-open if Terran is offline). Guard can also use a Terran exclusion namespace for concept protection. Buying Wire Professional does not include Terran.
Who it is for
VS Code developers
Want provider-agnostic chat in the editor with a local place to see every request, apply guardrails, and route or compress when it helps — spend estimates included — without a cloud account or admin rights.
Windows developers who want full control
Keep the VS Code extension for chat, and add the desktop command center when you
want session timeline, Sandbox, Privacy Audit, complete Settings, and the wire CLI — same pipeline and license.
Security-conscious users
Need allowlist-scoped inspection, secret and concept controls, and auditable proof that prompts are not written to disk — on your own machine.
One-time pricing
No subscriptions. Try everything for 7-day — every module unlocked, no card. Then keep Observe for $39 or buy Professional for $99. Checkout lives on the pricing page.
Frequently asked questions
Is there a free trial?
Yes. Every new install starts with a 7-day full-pipeline trial — Observe plus Guard+, Aura Router, Aura Compress, Terran RAG injection, and Advise Inject — with no card required. Observe is not free after the trial: keep it for $39, add individual modules, or buy Professional for the full pipeline. Installs before our modular pricing launch keep perpetual Observe access at no charge.
Does Aura Wire send my prompts to Halo Softworks LLC or any cloud service?
No. Aura Wire runs locally on loopback. Sanitized telemetry — provider, model, tokens, latency, cost — stays in a local SQLite database on your machine. Prompts, responses, and message text are never persisted.
What is Aura Advise?
Aura Advise is a prompting coach trained on your own traffic. It scores rolling prompt quality (clarity / structure / discipline), spots retry storms, repeated pasting, and vague prompts, and shows personalized advice cards under Pipeline → Aura Advise. Prompt text is read in memory only — only rule keys and counts are stored. Dashboard coaching is included with Observe and the trial; Inject mode (append a short coaching note to eligible chat responses) requires the Advise Inject module.
Can I see traffic from my laptop and desktop in one place?
Each Aura Wire install keeps its own local dashboard. Combining Live Traffic across machines is not offered in the current release.
What do I install?
The Aura Wire .zip from halosoftworks.com. Extract it, then install the .vsix from inside VS Code: Extensions → Install from VSIX. (The zip avoids Visual Studio’s installer opening by mistake.) It runs in VS Code and compatible forks, finds or starts the local Aura Wire Host on your machine, and chats through it — no admin rights, certificate, or driver. Later builds arrive through the chat Update button. The Host runs on Windows today; a macOS Host is in progress. The Windows installer adds the optional desktop command center and the wire CLI.
What is the desktop command center?
An optional Windows desktop app on the same local Host the extension uses. It expands what you can do beyond the curated in-editor boards: full Live Traffic and Session Timeline (including compare), Sandbox, Privacy Audit, the full custom-dashboard catalog, Working samples / History on pipeline stages, complete Settings (provider keys, licensing, retention, updates), and the wire CLI on your PATH. Chat and Observe still work in VS Code without it — the desktop app adds screens and capabilities, not a second Host.
Which editors work with Aura Wire?
Editor-tab chat works in VS Code 1.85+ and compatible forks (Insiders, VSCodium, Cursor, Windsurf) through the extension. Every chat from the extension, and every request from the wire CLI or an SDK pointed at the loopback /v1 door, runs through the same local pipeline.
Can I use Aura Wire without changing my AI provider API keys?
Yes. In the extension, pick a provider and model per chat; keys live locally in Aura Wire Settings. When Aura Router sends traffic to a different provider, keys for target providers are stored locally in Settings → Provider keys.
Will Aura Wire slow down my requests?
Observation adds minimal overhead. Aura Router cross-provider translation and Terran RAG injection add latency proportional to those operations; Terran uses a short timeout and skips injection if the REST server is unavailable.
What is Aura Guard and what are protected concepts?
Aura Guard scans ephemeral request content before other extensions run. It detects inline secrets (OpenAI, Anthropic, AWS, JSON api_key, and Bearer patterns) and optionally protected concepts — sensitive topics from a local corpus or a Terran exclusion namespace, including paraphrases. Modes: Warn/Relaxed (Observe and the trial), Redact/Standard, or Block/Strict (Guard+ module). Topics are never redacted — only warned or blocked. Matched secret values never appear in telemetry.
Do I need Aura Terran?
No. Observe, Guard, Router, Compress, and Advise work without it. Wire’s Terran pipeline stage is a bridge to Aura Terran, which remains a separate product you install locally. The Wire bridge entitlement is included with any paid Terran license and with Wire Professional — it is not sold as a separate Wire module. Enable the bridge when Terran is running Direct RAG REST (default http://127.0.0.1:41800) and you want ranked corpus chunks injected into Wire chat. You can also point Guard at a Terran exclusion namespace for concept protection.
How is Aura Wire different from Cursor's built-in usage dashboard?
Aura Wire sees traffic at the pipeline layer — every chat from the extension and every /v1 client — across providers and tools, with Guard, Router, Compress, and Advise actions, repeated-context analysis, local cost configuration, and a verifiable no-content-storage privacy audit. It is provider- and tool-agnostic, not Cursor-only.
What does Aura Wire actually store?
Only sanitized metadata: provider, model, endpoint, status, byte counts, token estimates, latency, anonymous Guard findings, Advise rule keys/counts, and HMAC fingerprints of repeated blocks. Never prompts, responses, tool arguments, or API keys.
What providers does Aura Wire support?
OpenAI, Azure OpenAI, OpenAI-compatible APIs, Anthropic, Google Gemini, local Ollama, and Cursor backend hosts when configured. Aura Router can translate requests and responses between providers automatically, including streaming and tool calls.
What are the safe first-run defaults?
During the trial or with an Observe license ($39 after trial — Observe is not free), Observe runs continuously and records sanitized metadata only. Guard, Router, Compress, Terran, and Advise start off until you enable them. When enabled, Guard defaults to Relaxed/Warn (detect secrets and concept similarity; never block or rewrite), Router to Suggest (recommend routes, do not rewrite), and Advise to Dashboard delivery (never inject until you choose). Compress and Terran stay off until you enable them.
What is the extension pipeline order?
Guard → Terran → Router → Compress → Observe. Guard runs first for block/redact/warn, then RAG injection, routing, compression, and finally sanitized telemetry is recorded. Aura Advise analyzes the same traffic in the observation path; Inject mode appends a short coaching note after Compress expansion when enabled.
Is HTTPS inspection required?
No. Editor-tab chat and /v1 clients talk to the local Host directly on loopback, so nothing is decrypted, no root certificate is installed, and no driver is required. Traffic leaves your machine to the provider you chose, over the provider’s normal HTTPS.
Does Aura Wire need administrator rights?
No. The VS Code extension installs from a .vsix without elevation, and the Host it finds or starts runs as your user. The Windows desktop app installs per-user as well. Aura Wire does not install a certificate, a driver, or a Windows service.
What platforms are supported?
The Aura Wire extension runs in VS Code and compatible forks. The local Host that backs chat is Windows today; a macOS Host is in progress. The desktop command center and wire CLI ship in the Windows installer, which auto-updates via Cloudflare R2.
Is Aura Wire a subscription?
No. Every install starts with a 7-day full-pipeline trial. After trial, keep Observe for $39 and add only the modules you need — or buy Professional for the full pipeline. Paid licenses are one-time purchases for the current major version. Future major upgrades may be paid, with discounts for existing paid customers. Installs from before the modular pricing launch keep perpetual Observe at no charge.
Why does Windows say the app is unrecognized?
Halo Softworks is a new software publisher, so Microsoft Defender SmartScreen may not yet have enough reputation data for every installer or release. Halo Softworks installers are digitally signed with Microsoft Azure Artifact Signing. Confirm that the publisher is Halo Softworks and that the digital signature is valid before continuing. The Secure Windows Downloads page on this site has step-by-step verification.
Is the installer digitally signed?
Yes. Halo Softworks signs Windows installers using Microsoft Azure Artifact Signing. You can verify the signature from the file's Properties dialog under Digital Signatures, or compare the SHA-512 checksum shown on the product download page.
Should I disable SmartScreen?
No. Halo Softworks does not recommend disabling SmartScreen. Verify the publisher, digital signature, download domain, and SHA-512 checksum instead. See Secure Windows Downloads (linked from Download and the site footer).
What should I do if the publisher is not Halo Softworks?
Do not run the installer. Delete it and download a fresh copy directly from the Halo Softworks website. Contact support if the problem continues.