From 042608720f8f0380266cd5ee9a315c3170c6b208 Mon Sep 17 00:00:00 2001 From: Sazed Bot Date: Mon, 25 May 2026 15:39:06 -0600 Subject: [PATCH] =?UTF-8?q?chore:=20SEO=20improvements=20=E2=80=94=20READM?= =?UTF-8?q?E=20opener=20+=20canonical=20LICENSE?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Restructure README opening to lead with concrete value (94+ recipes, vendor list, deliverables) before scope/policy. Improves GitHub search snippets and first-visitor conversion. - Add canonical `LICENSE` file (copy of LICENSE-CODE = MIT) so GitHub's classifier surfaces the MIT badge instead of NOASSERTION. Dual-license split (LICENSE-CODE / LICENSE-DATA / LICENSE.md) is preserved. - Link the `good first issue` label directly from the Contributing section to surface contribution paths. Part of a broader SEO/discoverability pass — also includes: - Description rewrite (manual via Settings, requires repo admin) - Topic list expansion to 20 tags (manual via Settings) - Enable Discussions (manual via Settings) --- LICENSE | 26 ++++++++++++++++++++++++++ README.md | 37 ++++++++++++++++++++++++------------- 2 files changed, 50 insertions(+), 13 deletions(-) create mode 100644 LICENSE diff --git a/LICENSE b/LICENSE new file mode 100644 index 0000000..88f6ce5 --- /dev/null +++ b/LICENSE @@ -0,0 +1,26 @@ +MIT License + +Copyright (c) 2026 Speechstack + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. + +This license applies to all source code in this repository, including but not +limited to: the JSON schema, validation scripts, GitHub Actions workflows, +and configuration files. Recipe data files in the `recipes/` directory are +licensed separately under CC-BY 4.0 (see LICENSE-DATA). diff --git a/README.md b/README.md index 0854352..bd23220 100644 --- a/README.md +++ b/README.md @@ -1,18 +1,29 @@ # Speechstack Recipes -> The open-source collection of production-ready recipes for building voice AI agents. +> 94+ production-ready voice AI agent stacks. Every recipe ships with real prompts, measured latency, and per-minute unit economics. -## Scope +🌐 **Browse the directory:** [speechstack.com](https://speechstack.com) -**SpeechStack covers voice + AI templates that combine 2+ tools for a specific outcome. We do not cover general AI tools.** +📊 **What's inside:** 94+ recipes · 10+ frameworks (Vapi, Pipecat, LiveKit, Retell, Bland, Vocode) · 8 STT vendors (Deepgram, AssemblyAI, Whisper, Google, Speechmatics…) · 7 TTS vendors (ElevenLabs, Cartesia, PlayHT, Azure, Resemble…) · 12+ industries + +## What's a recipe? + +A working voice agent documented as a single JSON file: -Every recipe in this repo must satisfy the [4-criteria template definition](./CONTRIBUTING.md#the-4-criteria-template-definition): schema-able artifact, forkable or copyable, specific named outcome, multi-component stack. PRs that don't clear the bar are closed, not merged "as drafts." +- **Framework** (Vapi / Pipecat / LiveKit / Retell / Bland / Vocode / …) +- **STT** + **LLM** + **TTS** + **Telephony** +- The actual system prompt (long-form Markdown) +- Measured latency (P50/P90 where known) +- Per-minute cost breakdown +- A link to public source code or a live demo ---- +If it doesn't ship all of that, it's a blog post link — not a recipe. -This repository powers [speechstack.com](https://speechstack.com) — a curated directory of real-world voice agent implementations across Vapi, Retell, LiveKit, Cartesia, ElevenLabs, Pipecat, Bland, and the rest of the voice AI stack. +## Scope + +**SpeechStack covers voice + AI templates that combine 2+ tools for a specific outcome. We do not cover general AI tools.** -Each recipe is a single JSON file that documents a working voice agent: its use case, the full stack (telephony + STT + LLM + TTS), unit economics, latency, prompt, configuration, and a link to a public source repo or demo. +Every recipe must satisfy the [4-criteria template definition](./CONTRIBUTING.md#the-4-criteria-template-definition): schema-able artifact, forkable or copyable, specific named outcome, multi-component stack. PRs that don't clear the bar are closed, not merged "as drafts." ## Contributing a recipe @@ -21,10 +32,10 @@ Got a voice agent you've built? We'd love to feature it. 1. Copy [`recipes/_template.json`](./recipes/_template.json) to `recipes/your-recipe-slug.json` 2. Fill in the fields (see [`schema/recipe.schema.json`](./schema/recipe.schema.json) for the spec) 3. Open a pull request -4. Our bot will validate it automatically in ~30 seconds +4. Our bot validates it automatically in ~30 seconds 5. We review and merge within 48 hours -See [CONTRIBUTING.md](./CONTRIBUTING.md) for the full guide. +See [CONTRIBUTING.md](./CONTRIBUTING.md) for the full guide. New contributors: check the [`good first issue`](https://github.com/speechstack-ai/recipes/labels/good%20first%20issue) label for recipes the community has specifically requested. ## What makes a good recipe @@ -68,7 +79,7 @@ npm install npm run validate ``` -This runs the same checks the CI does: +This runs the same checks CI does: - JSON schema validation - Recipe ID and slug uniqueness - Vendor name allowlist @@ -79,10 +90,10 @@ This runs the same checks the CI does: This repository uses a dual license: -- **Code** (schemas, validators, scripts, workflows) is licensed under [MIT](./LICENSE-CODE). -- **Recipe data** (everything in `recipes/`) is licensed under [CC-BY 4.0](./LICENSE-DATA). +- **Code** (schemas, validators, scripts, workflows) is [MIT](./LICENSE-CODE). +- **Recipe data** (everything in `recipes/`) is [CC-BY 4.0](./LICENSE-DATA). -In plain English: you can use the recipes in commercial work as long as you credit Speechstack. You can use the code however you want. See [LICENSE.md](./LICENSE.md) for the explainer. +Plain English: use the code however you want; use the recipes in commercial work as long as you credit Speechstack. See [LICENSE.md](./LICENSE.md) for the explainer. ## Stats