streamdouble 0.1.0__tar.gz → 0.1.0.dev0__tar.gz

{streamdouble-0.1.0 → streamdouble-0.1.0.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: streamdouble
-Version: 0.1.0
+Version: 0.1.0.dev0
 Summary: A framework-agnostic Twilio Media Streams simulator. Test your voice agent's WebSocket endpoint locally, at full protocol fidelity, without placing a real call.
 Project-URL: Homepage, https://github.com/shaswat28/streamdouble
 Project-URL: Issues, https://github.com/shaswat28/streamdouble/issues
@@ -40,10 +40,9 @@ WebSocket endpoint locally, at full protocol fidelity, without placing a real ca
 
 ![streamdouble placing a simulated Twilio call](docs/demo.svg)
 
-> **Status: early, but it works.** Places real calls, records the reply,
-> simulates bad networks, runs scripted scenarios, and gates on latency in CI.
-> Validated against a production voice agent, where it found a real bug.
-> Expect the API to move before 1.0.
+> **Status: pre-alpha, under active construction.** Phases 1–5 of 6 complete.
+> Places real calls, records the reply, simulates bad networks, runs scripted
+> scenarios, and gates on latency in CI. Not yet published to PyPI.
 
 ---
 
@@ -65,27 +64,12 @@ enough that your agent cannot tell it isn't Twilio.
 Two terminals. No Twilio account, no tunnel, no phone.
 
 ```bash
-pip install streamdouble
-```
-
-Point it at your agent, with any WAV as the caller's voice:
-
-```bash
-streamdouble call ws://localhost:8000/media-stream --audio hello.wav --out reply.wav
-```
-
-`reply.wav` is what your agent said back.
-
-**No voice agent yet?** Clone this repository — it ships one to talk to, plus
-the audio fixtures:
-
-```bash
-git clone https://github.com/shaswat28/streamdouble && cd streamdouble
-python examples/echo_agent.py --port 8000          # terminal one
+pip install -e ".[dev]"
+python examples/echo_agent.py --port 8000
 ```
 
 ```bash
-streamdouble call ws://localhost:8000/media-stream   --audio fixtures/speech_8k.wav --out reply.wav   # terminal two
+streamdouble call ws://localhost:8000/media-stream --audio fixtures/speech_8k.wav --out reply.wav
 ```
 
 ```
@@ -165,7 +149,7 @@ name: caller interrupts the greeting
 steps:
   - wait_for: audio        # let the agent get going
   - wait: 1.5              # streaming silence, not stopping
-  - say: ../fixtures/speech/interrupt.wav
+  - say: ../fixtures/speech_8k.wav
   - expect: clear          # a correct agent stops talking
   - wait: 1.0
 ```

{streamdouble-0.1.0 → streamdouble-0.1.0.dev0}/PLAN.md

@@ -3,10 +3,9 @@
 > A framework-agnostic Twilio Media Streams simulator. Test your voice agent's
 > WebSocket endpoint locally, at full protocol fidelity, without placing a real call.
 
-**Status:** All six phases complete, gates 1-4 passed. Public and published on
-PyPI. Gate 5's clean-install half is done -- the package was installed from
-PyPI into an empty virtualenv and exercised; what is still owed is someone who
-is not the author reading the README cold.
+**Status:** Phases 1-5 complete, gates 1-4 passed, Phase 6 under way. Public,
+not yet on PyPI. Gate 5 (fresh-eyes / clean-machine) is partly done and partly
+owed.
 **License:** Apache-2.0
 **Language:** Python 3.11+
 **Name:** `streamdouble` — `dialtone` was taken on PyPI. See [Progress](#progress).
@@ -342,9 +341,9 @@ That is the proof the tool has value beyond the happy path.
 - [x] CONTRIBUTING.md and issue templates. The protocol-difference template
       leads with "tell us where the simulation is wrong", because that is the
       most valuable report this project can receive.
-- [x] **PyPI publish — done.** `.github/workflows/release.yml` stays
-      manual-dispatch with `publish: false` by default, so no release happens
-      by accident. 0.1.0.dev0 went out first; 0.1.0 is the real one.
+- [ ] **PyPI publish — not yet done.** `.github/workflows/release.yml` is
+      manual-dispatch only, with `publish: false` by default, so the ordinary
+      run builds and verifies without uploading anything.
 
 > ### 🟡 REVIEW GATE 5 — Fresh-eyes pass — PARTLY DONE
 >
@@ -370,7 +369,7 @@ That is the proof the tool has value beyond the happy path.
 
 ---
 
-### Phase 6 — Distribution ✅ DONE
+### Phase 6 — Distribution 🔨 IN PROGRESS
 
 - [x] **Claude Code skill wrapper** — `skill/SKILL.md`. Carries the judgement
       as well as the commands: disable what the agent touches on hangup before
@@ -433,7 +432,7 @@ something personally useful; that is the checkpoint that sustains the rest.
 - [ ] A stranger installs it and gets a result in under five minutes
 - [ ] Reproduces at least one real bug that a phone call would have been needed to find
 - [ ] README has a demo GIF and an honest comparison table
-- [x] Published to PyPI under Apache-2.0
+- [ ] Published to PyPI under AGPL-3.0
 
 ---
 

{streamdouble-0.1.0 → streamdouble-0.1.0.dev0}/README.md

@@ -5,10 +5,9 @@ WebSocket endpoint locally, at full protocol fidelity, without placing a real ca
 
 ![streamdouble placing a simulated Twilio call](docs/demo.svg)
 
-> **Status: early, but it works.** Places real calls, records the reply,
-> simulates bad networks, runs scripted scenarios, and gates on latency in CI.
-> Validated against a production voice agent, where it found a real bug.
-> Expect the API to move before 1.0.
+> **Status: pre-alpha, under active construction.** Phases 1–5 of 6 complete.
+> Places real calls, records the reply, simulates bad networks, runs scripted
+> scenarios, and gates on latency in CI. Not yet published to PyPI.
 
 ---
 
@@ -30,27 +29,12 @@ enough that your agent cannot tell it isn't Twilio.
 Two terminals. No Twilio account, no tunnel, no phone.
 
 ```bash
-pip install streamdouble
-```
-
-Point it at your agent, with any WAV as the caller's voice:
-
-```bash
-streamdouble call ws://localhost:8000/media-stream --audio hello.wav --out reply.wav
-```
-
-`reply.wav` is what your agent said back.
-
-**No voice agent yet?** Clone this repository — it ships one to talk to, plus
-the audio fixtures:
-
-```bash
-git clone https://github.com/shaswat28/streamdouble && cd streamdouble
-python examples/echo_agent.py --port 8000          # terminal one
+pip install -e ".[dev]"
+python examples/echo_agent.py --port 8000
 ```
 
 ```bash
-streamdouble call ws://localhost:8000/media-stream   --audio fixtures/speech_8k.wav --out reply.wav   # terminal two
+streamdouble call ws://localhost:8000/media-stream --audio fixtures/speech_8k.wav --out reply.wav
 ```
 
 ```
@@ -130,7 +114,7 @@ name: caller interrupts the greeting
 steps:
   - wait_for: audio        # let the agent get going
   - wait: 1.5              # streaming silence, not stopping
-  - say: ../fixtures/speech/interrupt.wav
+  - say: ../fixtures/speech_8k.wav
   - expect: clear          # a correct agent stops talking
   - wait: 1.0
 ```

{streamdouble-0.1.0 → streamdouble-0.1.0.dev0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "streamdouble"
-version = "0.1.0"
+version = "0.1.0.dev0"
 description = "A framework-agnostic Twilio Media Streams simulator. Test your voice agent's WebSocket endpoint locally, at full protocol fidelity, without placing a real call."
 readme = "README.md"
 requires-python = ">=3.11"

{streamdouble-0.1.0 → streamdouble-0.1.0.dev0}/src/streamdouble/__init__.py

@@ -9,6 +9,6 @@ Twilio Media Streams protocol.
 
 from __future__ import annotations
 
-__version__ = "0.1.0"
+__version__ = "0.1.0.dev0"
 
 __all__ = ["__version__"]