2026-10-12 18:51 UTC

Swobu’s maintainers claim their released local switchboard pools hosted and local LLM capacity behind stable, shareable routes with per-request protocol translation and fallback, letting coding agents change providers without client reconfiguration or distributing provider credentials.

state: seedheat: lowuncertainty: highknownscott: lowlocal-inference llm-routing agent-harnessesSwobuswobuforge

What is this?

Swobu is a local-first LLM routing gateway published under the GitHub account swobuforge, with listings for VS Code, Claude Code and Codex. Its repository and product snippets describe keeping provider credentials at the gateway and exposing stable client routes while provider selection, balancing and fallback happen underneath. These first-party descriptions support the core routing claim, but the supplied snippets do not establish per-request protocol translation, E2E HTTPS sharing, or the specific pooling of hosted accounts, regions and local models; they also do not identify individual maintainers or establish release timing.

Why it matters to Scott

Swobu’s supported stable-route and backend-fallback claims repeat architecture Scott already operates in LiteLLM and the Ask terminal agent, consistent with Model Perishability; no supplied radar page tracks Swobu itself. The evidence does not establish consequential adoption or the potentially differentiating protocol-translation, encrypted-sharing and capacity-pooling claims, so it gives Scott no demonstrated reason to change his gateway or argument.
dev:technology.litellmdev:project.askip:concept.model-perishabilityradar:experiential-open-model-gatewayradar:concept.llm-routingradar:pangolin-wireguard-ai-gateway
queries asked of Scott's wikis
  • provider-independent agent harnesses stable model endpoints
  • multi-provider routing load balancing failover
  • local and hosted model capacity pooling
  • LLM protocol translation tool-calling compatibility
  • shared inference gateways credential isolation
  • model switching without client reconfiguration

Measured heat

now 0 pts/hpeak 0 pts/hcomments 0/hpeers p50momentum: steady2 platformsage 692h
points/hour across evidence · reading as of 2026-10-13 05:35:48.225472+11:00 · deterministic, not a model opinion

How the heat travelled

09-13 22:21 (minted)⭐ origin echo-reconstructedSwobu publishes a local LLM switchboard that pools provider accounts, regions, and local models behind stable routes and shares HTTPS endpoi
swobuforge on github (echo) · attributed from hn.story.49689250 · published time unknown
—
09-13 22:12first on hacker news · published · lag ?Show HN: Swobu – Local LLM Switchboard You Can Share over E2E HTTPS
metrofun
—
09-13 22:12amplified on hacker news 👑hn.story.49689250
metrofun
peak 3 · 1 comments · 65% of case engagement
09-18 16:48amplified on hacker newshn.story.49756982
metrofun
peak 2 · 0 comments · 33% of case engagement
09-13 22:20our radar first saw it · lag ?discovery anchor: hn.story.49689250—
pace: p35 vs 1037 stories at the 336h mark (now 692h old) — ahead of agentgate-signed-agent-receipts (1.3x), behind agent-memory-add-search-evaluation (0.8x)

Evidence (3) — ⭐ canonical anchor

sourceobjectauthorscorecomments
🟧 hnShow HN: Swobu – Local LLM Switchboard You Can Share over E2E HTTPS
Retrieved article excerpt

Open article · Retrieved 2026-09-13T22:21:41.075622+00:00

# [Swobu](https://swobu.com/) — LLM Switchboard

**Pool the LLM capacity you already have.**

Put AWS/Azure credits, provider accounts and quota, free APIs and local GPUs behind stable routes for Claude Code, Codex and other agents. Change the capacity behind a route without reconfiguring every client.

[Documentation](https://swobu.com/docs/) · [Quickstart](https://swobu.com/docs/start/first-route/) · [Releases](https://github.com/swobuforge/swobu/releases) · [VS Code](https://marketplace.visualstudio.com/items?itemName=swobu.swobu)

```
curl -fsSL https://swobu.com/install.sh | sh
```

Windows PowerShell: `irm https://swobu.com/install.ps1 | iex`

[Claude Code using one Swobu route across Azure, Bedrock, Gemini and OpenAI](https://github.com/swobuforge/swobu/blob/master/assets/readme/swobu-demo.gif)

## One route. Your capacity underneath.

A Swobu route looks like a model name to the client. You decide what sits behind it.

```
Claude Code / Codex / OpenCode
              │
              │  model: work
              ▼
            Swobu
              │
              ├─ Azure / westus2 / gpt-5.6-sol
              ├─ Azure / westcentralus / gpt-5.6-sol
              └─ fallback: OpenAI / gpt-5.6-sol
```

The client keeps asking for `work`. You can change accounts, regions, providers or models behind that route independently.

**Runtime rule:** Swobu does not preflight compatibility. It attempts the real request against the configured target. If that attempt fails, it tries the next target in the route. Protocol translation happens per attempt; providers remain different.

## What the switchboard is for

| Job | Route pattern |
| --- | --- |
| **Use funded capacity before cash** | AWS/Azure/GCP-backed targets before direct paid APIs |
| **Aggregate quota** | same model across regions or accounts |
| **Make cheap capacity useful** | free/local targets first, paid fallback behind them |
| **Route agents by workload** | premium main route, cheap/free worker routes |
| **Keep policy out of clients** | local-first, region/provider restrictions, funded-first ordering |
| **Share configured capacity** | one HTTPS endpoint while routing and provider credentials stay owner-side |

Swobu does not inspect your credit balance, remaining TPM, price or latency. The route expresses your policy; Swobu executes it.

## One client endpoint, multiple provider protocols

```
Claude Code ─┐
Codex ───────┤
OpenCode ────┤
Kilo ────────┤
Pi ──────────┼──── Swobu ────┬─ OpenAI Responses / Chat Completions
OpenClaw ────┤                ├─ Anthropic Messages
Hermes ──────┘                ├─ Gemini
                              └─ provider-specific endpoints
```

Swobu translates supported request semantics for each target attempt. Translation is not equivalence: provider-specific behavior remains observable, and unsupported combinations can fail and advance through the route.

Current docs cover **40 provider integrations**, including local inference, frontier APIs, hyperscalers and specialized inference platforms.

[Clients and provider setup →](https://swobu.com/docs/)

## Share the endpoint. Keep provider keys at home.

```
swobu share dev/coding
```

A shared route gives the recipient an HTTPS endpoint and Swobu bearer. Routing, fallback and provider credentials remain on the owner side. Change the targets later without changing the recipient's endpoint or route name.

[Swobu Shared API page showing OpenAI-compatible and Anthropic-compatible endpoints](https://github.com/swobuforge/swobu/blob/master/assets/readme/shared-api.png)

[Workspace and Route Share →](https://swobu.com/docs/concepts/sharing/)

## Connect an agent

Cockpit can configure supported clients, or use the CLI:

```
swobu connect claude
swobu connect codex
swobu connect opencode
swobu connect kilo
swobu connect pi
```

Provider credentials and routing stay behind Swobu rather than being copied into every client.

## Local-first

Swobu runs locally and exposes the endpoint your agents connect to. No Swobu account is required for local use.

Swobu sends privacy-minimized usage and reliability telemetry tied to a random installation ID. It does **not** send prompts, responses, credentials, endpoints or user-defined names. Disable it with `swobu telemetry off` or `DO_NOT_TRACK`.

[Security and privacy →](https://swobu.com/docs/)

## Build from source

```
git clone https://github.com/swobuforge/swobu.git
cd swobu
make build
./.out/swobu --version
```

Swobu publishes versioned binaries for Linux, macOS and Windows with SHA-256 checksums.

[Latest release →](https://github.com/swobuforge/swobu/releases/latest)

## License

Swobu is available under the [GNU AGPLv3](https://github.com/swobuforge/swobu/blob/master/LICENSE). A [commercial license](https://github.com/swobuforge/swobu/blob/master/COMMERCIAL-LICENSE.md) is also available.

---

**Translations:** [English](https://github.com/swobuforge/swobu/blob/master/README.md) · [简体中文](https://github.com/swobuforge/swobu/blob/master/README.zh-CN.md) · [日本語](https://github.com/swobuforge/swobu/blob/master/README.ja.md) · [Português (Brasil)](https://github.com/swobuforge/swobu/blob/master/README.pt-BR.md) · [Bahasa Indonesia](https://github.com/swobuforge/swobu/blob/master/README.id.md) · [한국어](https://github.com/swobuforge/swobu/blob/master/README.ko.md) · [Русский](https://github.com/swobuforge/swobu/blob/master/README.ru.md) · [Español](https://github.com/swobuforge/swobu/blob/master/README.es.md) · [Українська](https://github.com/swobuforge/swobu/blob/master/README.uk.md)
metrofun31
🟧 echo.github ⭐Swobu publishes a local LLM switchboard that pools provider accounts, regions, and local models behind stable routes and shares HTTPS endpoiswobuforge——
🟧 hnShare an LLM endpoint over HTTPS while TLS still terminates on your machinemetrofun20

Interpretation history

Decision trace