Friday
A local control plane that routes AI work with approvals, verification, and audit evidence.
- Source repo
- thesongzhu/Friday
- Stars
- ★ 833
- Last updated
- 3mo ago
- License
- MIT
- Primary language
- TypeScript
- FA score
- 66/100 · Some gaps
At a glance
- How it runs
- Works with
- Universal · cross-platformCodex · Claude.ai
- Cost
- Free software; you pay for model usage
- Setup effort
- Medium · a few setup steps
- You'll need
- Typical use
- A solo developer using several model APIs can hand Friday a long-running software goal and inspect or override the model chosen for each step.
- Not a fit if
- Users who want a ready-made consumer installer: no public native Friday.app yet
- Teams needing production-proven agents, not internal self-test traffic
- Source review
- 66/100 · Some gaps
What does this agent do, and when should you use it?
Friday is a private control plane for AI agents, not a model or chatbot. Its controller, keys, data, and memory are intended to live on the local machine, while BYOK connects to cloud AIs such as Codex, Claude, and DeepSeek. Its stated loop is route, execute, verify, and remember: select a model, govern risky work through approval, require evidence of completion, and propose memories for confirmation. The README describes hash-chained audit records for routed turns, including actor, content, and metering information. The current product runs from source; the npm package is explicitly not a complete consumer installation path, and the native Friday.app is not publicly released.
After a user supplies a goal, Friday follows the documented route → execute → verify → remember loop. It selects an AI for a step, executes governed work, and puts risky actions into an approval-or-rejection path. Completion requires a proof receipt rather than a model's assertion alone; the README describes this as enforced through a persisted mission/work-item code boundary. It presents memories as candidates that require user confirmation before becoming stored memories, and says recalled content is scrubbed for known secret patterns before being shown to a model. Routed agent turns are described as all-or-nothing entries in a hash-chained audit log with execution and cost details.
- A solo developer using several model APIs can hand Friday a long-running software goal and inspect or override the model chosen for each step.
- A technical lead automating operational work can allow ordinary read-only actions while requiring a separate approval for actions that change external state.
- A team that needs stronger evidence for agent output can require a proof receipt before a work item is treated as complete.
- A BYOK user managing spend across providers can route work before execution and retain a record of the model that actually answered after a fallback.
- A privacy-conscious individual can review each proposed memory candidate before allowing it to influence later work.
How do you install or deploy this agent?
The README provides this macOS source-run path:
brew install node@22
git clone https://github.com/thesongzhu/Friday.git
cd Friday
open "Friday Setup.command"It requires Node.js 22 or later and your own model-provider keys. The supplied evidence does not document a first CLI invocation, credential configuration format, or a non-macOS installation path. It also explicitly states that npm install does not install the full Friday product.
How do you use this agent?
Run "Friday Setup.command" after the documented source checkout. Then submit a goal through the described chat interface; Friday processes it through route, execute, verify, and remember, pausing risky steps for approval. You must provide your own keys for the named cloud AIs, including Codex, Claude, or DeepSeek. The supplied evidence does not document a copyable first-goal command, chat endpoint, or configuration-file format.
What are this agent's strengths and limitations?
- Keeps the controller, keys, data, and memory on the local machine while supporting BYOK cloud models from multiple vendors.
- Separates risky actions into human approval instead of allowing an agent to approve its own side effects.
- Treats proof receipts as the basis for task completion instead of relying only on an agent's self-report.
- Describes routing, metering, and hash-chained audit records for tracing execution and model cost.
- It currently runs from source; there is no publicly released native Friday.app or documented consumer installer.
- The source tree is 1.0.3 while the npm package is 1.0.2, and the README says the package does not carry the complete current engine.
- It requires Node.js 22+, a local runtime environment, and user-supplied cloud-model keys rather than local model execution.
- The README says live behavior has so far run only on internal and self-test traffic, not real end-user workloads.
- More autonomous capabilities are off by default, while capability acquisition and self-upgrade remain review-gated work in progress.
How does this agent compare with similar options?
Friday positions itself as a local, cross-vendor control layer: Codex and Claude are presented as individual agent providers, while Friday handles routing, approval, verification, memory-candidate confirmation, and auditing across them.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| Friday This agent | 66 · Some gaps | CLIFree + model costs | ★ 833 | 3mo ago | TypeScript | Codex · Claude.ai |
| taOS Self-Hosted Agent OS | 68 · Some gaps | Self-hosted serviceFree | ★ 604 | 14d ago | Python | Claude Code · OpenAI API · Claude API |
| Agents Squads | 63 · Some gaps | CLIFree + model costs | ★ 56 | 19d ago | TypeScript | Claude Code |
| Manifold | 55 · Major gaps | Self-hosted serviceFree | ★ 503 | 17d ago | Go | OpenAI API · Claude API |
How does FollowAgents rate this agent?
Why each dimension lost points
The README describes local control, BYOK, approval for risky steps, confirmation before promoting memory candidates, audit records, and proof-bearing completion. These give least privilege, confirmation, data-flow transparency, and sensitive-data handling substantive design support, but the packet is mostly product description and the supplied code does not show the full implementations, so these do not earn full marks. CI pins some GitHub Actions to commit hashes, and the docs discuss encryption and key protection, but dependency auditing and vulnerability-response evidence is limited. Approval paths for external effects are described, while rollback and recovery after revocation are thin. The contributor copyright notice offers limited attribution; publisher identity is unverified and ongoing ownership is unclear.
The README, Getting Started guide, and package metadata are broadly consistent about Node 22, source-based operation, and the npm package not being the consumer product. The docs also give health checks, troubleshooting steps, and capability-verification guidance. The RFC distinguishes shipped scope from default-off status, reducing the risk of treating design goals as current behavior. The packet is incomplete and cannot establish runtime behavior, so consistency, dependency availability, and failure-message scores remain moderate.
The docs identify developers, maintainers, and CI as audiences and describe provider, channel, skill, and task scenarios. They also state concrete limits, including the unreleased native app, disabled autonomy features, and external-account blockers, supporting high scores here. Trigger phrases appear only for the source-validation skill; the other sampled skills provide no trigger-precision or false-trigger guidance. Node 22, source, Docker, and OS-specific state paths show some environment fit, but the quickstart is macOS-oriented and coverage is incomplete.
The README, setup guide, troubleshooting guide, architecture RFC, package manifest, and LICENSE provide a recognizable information structure. Installation paths, prerequisites, and npm-package limits are explained well; the license is explicitly MIT. Skill examples and troubleshooting are useful, though FAQ coverage is limited. Package version and release status are stated, but the packet does not include a complete change history. The RFC names the Platform Team and the license credits Friday contributors, but no ongoing maintainer or security-response owner is clearly identified.
The product targets cross-model routing, governance, verification, and audit. The README explains a work loop and claimed differentiators, while the setup guide specifies a useful response shape for capability status and next steps. These support operationally usable output. The proposed marginal value is clear, but the packet contains no user-workflow or outcome evidence demonstrating the benefit. Models, search, and third-party services require user-supplied credentials and may incur costs; metering is mentioned without enough cost-control detail, so cost-benefit support is limited.
Several claims have corresponding evidence across the README, Getting Started guide, package manifest, RFC, and CI configuration, including the Node version, distribution limits, default-off multi-tenant feature, and CI gates. The README distinguishes current source, the older npm package, and future direction. The underlying implementations and specific test results for security guarantees are not included, so they are not independently established by this packet. Documentation labels drafts, default-off features, and unreleased plans, helping separate current facts from plans, though some strong guarantees remain assertions not sufficiently corroborated by the sample.
- This is a low-confidence static review. The packet omits full implementations for the stated security guarantees and cannot establish runtime behavior.
- The README says the npm package is not the consumer install path and the native Friday.app is not publicly released; multi-tenant support is default-off and some autonomy features are not enabled.
- Real tasks require user-configured model or third-party credentials and may involve external data transfer and costs; verify providers, permissions, and approval settings as documented.
Maintain this project? Add the FARS badge to your README
[](https://followagents.com/en/agents/friday)<a href="https://followagents.com/en/agents/friday"><img src="https://followagents.com/badge/friday.svg" alt="FARS review"></a>Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
Release notes excerpt
Friday 1.0.2 — public v1 local candidate hardening patch Release Claim Final release SHA: 68de607b17214db0fdf400c29b8bbe0876eb21fb Same-SHA R5 proof: Real Green Gate workflow_dispatch run 26469316019 — status=passed, scenarios 94/94, blocked_reasons=[], evidence_kinds_observed includes real-runtime, real-provider, real-browser, and manual-external. Discord, Telegram, and Lark/Feishu trusted-inbound channel artifacts all validate valid:true against the release SHA. …View on GitHub ↗ -
Release notes excerpt
Friday 1.0.1 — public v1 local candidate (npm/source-only distribution) Release Claim Final release SHA: 222bc6ce43431b4d51366c4deb0602a0c68b8a4e Same-SHA R5 proof: Real Green Gate workflow_dispatch run [26434697997](https://github.com/thesongzhu/Friday/actions/runs/26434697997) — status=passed, scenarios 94/94, blocked_reasons=[], evidence_kinds_observed includes real-runtime + real-provider + real-browser + manual-external. All three phase24X channel artifacts (Discord phase24b, Telegram phase24c, Lark/Feishu phase24d) status=passed + failures=[] + `envi …View on GitHub ↗ -
Release notes excerpt
What's Changed * [phase0] Automation Bootstrap And Guardrails by @thesongzhu in https://github.com/thesongzhu/Friday/pull/22 * Phase 1: enrich skills foundation surfaces by @thesongzhu in https://github.com/thesongzhu/Friday/pull/23 * Phase 2: add plugin SDK preview policy summaries by @thesongzhu in https://github.com/thesongzhu/Friday/pull/24 * Phase 3: add channel contract starter skills by @thesongzhu in https://github.com/thesongzhu/Friday/pull/25 * Phase 4: add fleet and browser diagnostics summaries by @thesongzhu in https://github.com/thesongzhu/Friday/pull/26 …View on GitHub ↗
FAQ
Does Friday include a model or model usage costs?
Will risky actions run automatically?
Can an agent mark work complete just by saying it is done?
Can I install the complete product from npm?
Thanks — we use this to improve the page.