BitRun

# Fix list: BitRun

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/bitrun, the October 2026 research run, assessed 15 October 2026. Grade F, 37.9 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on BitRun: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Security & auth, 28 out of 100, up to 12.6 more on the total

Why it scored 28: Read with the tool checklist, credential model first. No API key. BitRun doesn't pass llama-server's `--api-key`, so any local process can call the server. The bind address is fixed at 127.0.0.1 with no flag to change it (5). Inference only, with no tools, file access or write actions behind the API (10). The server returns model output and fetches no outside content (10). Engine logs go to `~/.bitrun/engine.log`. The server is started with `--log-disable`, and there is no per-call log (3). No SECURITY.md, security.txt or advisories. Model files from Hugging Face and release archives are fetched with no checksum check, and the install route is `curl | sh` (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 2. Reliability, 38 out of 100, up to 12.4 more on the total

Why it scored 38: Read with the local-software lines, since BitRun runs on the owner's machine with no hosted service. Archives for macOS arm64, Linux amd64 and Windows amd64 on GitHub Releases and one-line install scripts, with supported platforms stated in the README. No package registry, no checksums published, and the shell installer appends to the user's shell profile (15). No CI workflow (the repository has no .github folder). Three Go test files cover Modelfiles, the Cockpit view and power estimates, with nothing running them in public (0). No issue or pull request has been filed in the project's three days, so there is no record to judge. We gave the middle of the scale (15). Semver-style pre-release tags with notes on each release, but no changelog file, and the v0.2.0-beta engine swap from bitnet.cpp to patched llama.cpp is described as an improvement, not a breaking change (8). Version 0.2.0-beta, not 1.0 or declared stable (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 3. Schema & documentation, 30 out of 100, up to 11.4 more on the total

Why it scored 30: Read with the API lines, for the local server an agent calls. No spec of its own. The server is llama-server's OpenAI-compatible surface, and the README documents one route, `/v1/chat/completions`, with one curl and one Python example (3 of 25). No llms.txt or docs site. The README is Markdown readable raw on GitHub (3 of 10). The README states what each command does and which model suits which job, but nothing on when not to use the server (6). No parameter reference. The model tags are a fixed list in the README (4). Examples for curl and the OpenAI Python SDK. No error responses documented (6). Release notes per tag on GitHub, and the routes sit under `/v1` (8).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 4. Agent ergonomics, 40 out of 100, up to 9.8 more on the total

Why it scored 40: Read with the API lines. Output size is controlled through the OpenAI fields llama-server accepts, which BitRun doesn't document. Context windows run from 768 to 8,192 tokens by model (10). No list endpoints worth paging. One model per server process (5). Errors are whatever llama-server returns, with none documented by BitRun (5). Generation has no side effects, so a retry is safe, though nothing documents it (10). Few required parameters, a default port and a placeholder key, and OpenAI's SDKs work against it. No SDK of its own (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 5. Payments & pricing, 60 out of 100, up to 5 more on the total

Why it scored 60: Read with the self-hosted rule. No x402, MPP or L402, and nothing is sold (0). Free under MIT with no account, key or card, so 20, 20 and 20 on the last three lines.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 6. Transparency & trust, 50 out of 100, up to 4.4 more on the total

Made of editorial 46, provenance 53.

Why it scored 50: The editorial half. MIT, all source public, including the C++ patches (30). No privacy statement or statement of what the software sends. The Go source has no network telemetry, and the outbound calls we found are model downloads from huggingface.co on `pull` (8). No deprecation policy, and v0.2.0-beta replaced the engine within three days (0). The README's "telemetry" is local power and CPU readings shown on screen. Nothing says that nothing leaves the machine, and nothing needs an opt-out (8).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Domain age: , no registry record we could read (0 of 15)
- Status page: not found (0 of 10)
- security.txt: not found (0 of 10)

## 7. Maintenance & community, 59 out of 100, up to 3.6 more on the total

Why it scored 59: v0.2.0-beta on 15 October 2026 (30). Three releases in four days, 12 to 15 October 2026 (20). No issues or pull requests filed yet, so responsiveness can't be shown. Project only three days old (5). Not in the official MCP registry (no MCP server) and no SDK (0). Go 1.24 with current Bubble Tea dependencies, but no CI or dependency updates set up (4).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## Deductions

Each comes off the total. A fixed and documented problem counts for less at the next check.

- Misleading claim, 15 October 2026. The README's "Verified Hardware Benchmarks", described as physical hardware runs, list package power of 11.70 W, 22.60 W and 8.50 W for every model on each machine, and joules per token from them. These equal the fixed estimates in power.go (2.5 + 1.15 per thread on macOS without battery data, 5.0 + 2.2 per thread on Windows, 4.5 + 2.0 per thread on Linux without RAPL) at the default thread counts of 8, 8 and 2. The README also says Windows power is measured through PPM and that `pull` resumes downloads, and the source does neither (https://github.com/IshekKhal/bitrun/blob/0fec65a/pkg/telemetry/power.go#L137-L163, https://github.com/IshekKhal/bitrun/blob/0fec65a/pkg/models/downloader.go#L57-L60, https://github.com/IshekKhal/bitrun/blob/0fec65a/README.md). Judgement call, -3.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- The lead matched. BitRun by IshekKhal is a CPU runner for 1-bit and 1.58-bit models with an OpenAI-compatible API, v0.1.0-alpha on 12 October 2026. The Show HN post is dated 13 October 2026 in UTC (14 October in India, the commits' time zone).
- Judgement call on the deduction of 3 for the benchmark power figures and the resume claim. The arithmetic matches to the hundredth of a watt at the default thread counts, but we didn't run the software.
- We did not install or run BitRun. Every claim about behaviour comes from the source at commit 0fec65a, the README and the release notes.
- Which llama-server routes beyond `/v1/chat/completions` work in BitRun's patched build was not established. The patch label says b1395 while the patched ggml has recent types such as NVFP4, so the exact upstream base is unclear.
- unchecked: whether the release archives match the source, as no checksums or build provenance are published.

## Weaknesses

- First released on 12 October 2026. Three pre-release versions in four days, one maintainer and no CI workflow
- The README's package power and joules-per-token figures match a fixed per-thread formula in the code, not a counter reading
- No API key option, so any local process can call the server, and model downloads and the installer verify no checksum
- The server documentation covers one route, `/v1/chat/completions`, with no error responses, OpenAPI file or llms.txt
- Context windows of the catalogue models run from 768 to 8,192 tokens, and one server process serves one model

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Ask the user to install BitRun and run `bitrun pull <tag>` and `bitrun serve <tag>` first. Then call http://127.0.0.1:11435/v1 with any API key string
- Poll `/health` before the first request. The server loads the model in the background and answers only when ready
- Use a model tag from the catalogue (bitnet-70m, bonsai-1.7b, falcon3-1b, bitnet-2b, falcon3-3b, bonsai-8b) and keep prompts within its context window
- Treat the README's energy figures as estimates, and don't report them as measured power
- Expect small research checkpoints. Check outputs before relying on them for code or facts

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.