--- title: "sherpa-onnx vs TensorFlowASR" type: "comparison" canonical_url: "https://www.graphcanon.com/compare/k2-fsa-sherpa-onnx-vs-tensorspeech-tensorflowasr" tools: ["k2-fsa-sherpa-onnx", "tensorspeech-tensorflowasr"] --- # sherpa-onnx vs TensorFlowASR *GraphCanon updated Jul 31, 2026* ## Verdict Pick sherpa-onnx if sherpa-onnx offers local speech-to-text processing across multiple platforms including embedded systems with NPUs, supporting wide programming languages; pick TensorFlowASR if tensorFlowASR is an advanced Automatic Speech Recognition library built on TensorFlow 2 that supports modern architectures such as Conformer and RNN-Transducer. [sherpa-onnx](https://k2-fsa.github.io/sherpa/onnx/index.html) reports 14k GitHub stars, 1.6k forks, and 617 open issues, last pushed Jul 29, 2026. [TensorFlowASR](https://huylenguyen.com/asr) has 1.0k stars, 239 forks, and 47 open issues, last pushed Jul 30, 2026. Figures are from public GitHub metadata via [sherpa-onnx's repository](https://github.com/k2-fsa/sherpa-onnx) and [TensorFlowASR's repository](https://github.com/TensorSpeech/TensorFlowASR). | | [sherpa-onnx](/tools/k2-fsa-sherpa-onnx.md) | [TensorFlowASR](/tools/tensorspeech-tensorflowasr.md) | | --- | --- | --- | | Tagline | Speech-to-text and related audio processing tools using ONNX with cross-platform support | Almost State-of-the-art Automatic Speech Recognition in Tensorflow 2 | | Stars | 13,842 | 1,010 | | Forks | 1,596 | 239 | | Open issues | 617 | 47 | | Language | C++ | Python | | Adopt for | sherpa-onnx offers local speech-to-text processing across multiple platforms including embedded systems with NPUs, supporting wide programming languages. | TensorFlowASR is an advanced Automatic Speech Recognition library built on TensorFlow 2 that supports modern architectures such as Conformer and RNN-Transducer. | | Persona | - | - | | Runtime | - | - | | License | Apache-2.0 | Apache-2.0 | | Categories | Speech & Audio | Speech & Audio | ## Trust and health _Sourced signals - not a safety guarantee. No winner column._ | | [sherpa-onnx](/tools/k2-fsa-sherpa-onnx.md) | [TensorFlowASR](/tools/tensorspeech-tensorflowasr.md) | | --- | --- | --- | | Open issues (now) | 617 | 47 | | Full report | [trust report](/tools/k2-fsa-sherpa-onnx/trust.md) | [trust report](/tools/tensorspeech-tensorflowasr/trust.md) | ## Shared compatibility - **Python**: [sherpa-onnx](/tools/k2-fsa-sherpa-onnx.md) - Python runtime; [TensorFlowASR](/tools/tensorspeech-tensorflowasr.md) - Python runtime ## Decision facts: sherpa-onnx - **Adopt for:** sherpa-onnx offers local speech-to-text processing across multiple platforms including embedded systems with NPUs, supporting wide programming languages. ## Decision facts: TensorFlowASR - **Adopt for:** TensorFlowASR is an advanced Automatic Speech Recognition library built on TensorFlow 2 that supports modern architectures such as Conformer and RNN-Transducer. - **License detail:** Apache-2.0 ## Choose when ### Choose sherpa-onnx if… - sherpa-onnx is primarily C++; TensorFlowASR is Python. - Tags unique to sherpa-onnx: aarch64, android, arm32, asr. - When you need to deploy speech recognition or synthesis on platforms such as Raspberry Pi, RISC-V, and with NPUs like Rockchip, Axera, Ascend ### Choose TensorFlowASR if… - TensorFlowASR is primarily Python; sherpa-onnx is C++. - Tags unique to TensorFlowASR: automatic-speech-recognition, conformer, contextnet, ctc. - TensorFlowASR ships Docker support for self-hosted deployment. - When focusing on state-of-the-art performance with models like Conformer or ContextNet, which are among its supported architectures and not necessarily included in all ASR libraries. ## When NOT to use sherpa-onnx - If your project strictly requires real-time web service interaction or server-based speech-to-text solutions - When the target environment lacks support for ONNX runtime and specific NPUs integrated within sherpa-onnx ## When NOT to use TensorFlowASR - When Python version constraints are an issue, as TensorFlowASR strictly requires python >= 3.12 on Apple Silicon devices, limiting compatibility. - If your project cannot accommodate the extra installation steps involving `ctc_decoders` and `rnnt_loss`, which implies additional complexity in setting up the environment. ## Common questions ### What is the difference between sherpa-onnx and TensorFlowASR? sherpa-onnx: Speech-to-text and related audio processing tools using ONNX with cross-platform support. TensorFlowASR: Almost State-of-the-art Automatic Speech Recognition in Tensorflow 2. See the comparison table for live GitHub stats and shared categories. ### When should I choose sherpa-onnx over TensorFlowASR? Choose sherpa-onnx over TensorFlowASR when sherpa-onnx is primarily C++; TensorFlowASR is Python; Tags unique to sherpa-onnx: aarch64, android, arm32, asr; When you need to deploy speech recognition or synthesis on platforms such as Raspberry Pi, RISC-V, and with NPUs like Rockchip, Axera, Ascend. ### When should I choose TensorFlowASR over sherpa-onnx? Choose TensorFlowASR over sherpa-onnx when TensorFlowASR is primarily Python; sherpa-onnx is C++; Tags unique to TensorFlowASR: automatic-speech-recognition, conformer, contextnet, ctc; TensorFlowASR ships Docker support for self-hosted deployment; When focusing on state-of-the-art performance with models like Conformer or ContextNet, which are among its supported architectures and not necessarily included in all ASR libraries. ### When should I avoid sherpa-onnx? If your project strictly requires real-time web service interaction or server-based speech-to-text solutions When the target environment lacks support for ONNX runtime and specific NPUs integrated within sherpa-onnx ### When should I avoid TensorFlowASR? When Python version constraints are an issue, as TensorFlowASR strictly requires python >= 3.12 on Apple Silicon devices, limiting compatibility. If your project cannot accommodate the extra installation steps involving `ctc_decoders` and `rnnt_loss`, which implies additional complexity in setting up the environment. ### Is sherpa-onnx or TensorFlowASR more popular on GitHub? sherpa-onnx has more GitHub stars (13,842 vs 1,010). Stars measure visibility, not whether either tool fits your constraints. ### Are sherpa-onnx and TensorFlowASR open source? Yes - both are open-source projects on GitHub (sherpa-onnx: Apache-2.0, TensorFlowASR: Apache-2.0). ### Where can I find alternatives to sherpa-onnx or TensorFlowASR? GraphCanon lists graph-backed alternatives at [sherpa-onnx alternatives](/tools/k2-fsa-sherpa-onnx/alternatives) and [TensorFlowASR alternatives](/tools/tensorspeech-tensorflowasr/alternatives) ([sherpa-onnx markdown twin](/tools/k2-fsa-sherpa-onnx/alternatives.md), [TensorFlowASR markdown twin](/tools/tensorspeech-tensorflowasr/alternatives.md)), ranked by typed relationship edges rather than popularity votes. ### Is there a machine-readable version of this comparison? Yes. The markdown twin at [this comparison](/compare/k2-fsa-sherpa-onnx-vs-tensorspeech-tensorflowasr.md) mirrors this page for agents and LLM crawlers, with the same stats table and FAQ answers. ### Which is better maintained, sherpa-onnx or TensorFlowASR? sherpa-onnx: Very active. TensorFlowASR: Very active. Compare maintenance labels, days since push, and release cadence in the trust section below - stars alone do not measure maintenance. ### Where are the full trust reports for sherpa-onnx and TensorFlowASR? GraphCanon publishes per-repo trust reports with dated maintenance, provenance, and scan summaries: [sherpa-onnx trust report](/tools/k2-fsa-sherpa-onnx/trust); [TensorFlowASR trust report](/tools/tensorspeech-tensorflowasr/trust). --- **Machine-readable endpoints** - JSON: [`/api/graphcanon/graph?tool=k2-fsa-sherpa-onnx`](/api/graphcanon/graph?tool=k2-fsa-sherpa-onnx) - LLM index: [/llms.txt](/llms.txt) - Full corpus: [/llms-full.txt](/llms-full.txt) _GraphCanon - The knowledge graph for AI development. https://www.graphcanon.com/_